Datasets:
Edition 2026-11-12-e97c02573b96
Browse files- DATASHEET.md +1 -1
- README.md +1 -1
- adyt.csv +8 -8
- adyt.json +0 -0
- chanticleer.csv +0 -0
- chanticleer.json +2 -2
- datapackage.json +26 -26
- for-clinical-use.csv +13 -13
- for-procurement.csv +1 -1
- manifest.json +0 -0
- profiles.json +2 -2
- querist.csv +0 -0
- querist.json +2 -2
- records.csv +0 -0
- records.json +2 -2
- ro-crate-metadata.json +27 -27
- viator.csv +0 -0
- viator.json +2 -2
DATASHEET.md
CHANGED
|
@@ -32,7 +32,7 @@ The benefits can be filtered by sector, place, beneficiary and grade. Making the
|
|
| 32 |
|
| 33 |
A count of benefits is not a count of people. Since the AIBD contains only benefits, it has no denominator: Nobody can infer from it how often AI succeeds, what AI is worth in total, or whether the goods of AI outweigh its harms. Nor were the benefits sampled uniformly: Since they come from published sources, a benefit that nobody wrote about is missing.
|
| 34 |
|
| 35 |
-
Of the 2,190 benefits, 1,332 are supported only by their own parties. Of the 1,
|
| 36 |
|
| 37 |
## Distribution
|
| 38 |
|
|
|
|
| 32 |
|
| 33 |
A count of benefits is not a count of people. Since the AIBD contains only benefits, it has no denominator: Nobody can infer from it how often AI succeeds, what AI is worth in total, or whether the goods of AI outweigh its harms. Nor were the benefits sampled uniformly: Since they come from published sources, a benefit that nobody wrote about is missing.
|
| 34 |
|
| 35 |
+
Of the 2,190 benefits, 1,332 are supported only by their own parties. Of the 1,857 benefits whose place is known, 808 happened in the United States, China or India. Health is the largest sector, with 531 benefits. Of the 5,973 sources, 808 are quoted in translation from another language.
|
| 36 |
|
| 37 |
## Distribution
|
| 38 |
|
README.md
CHANGED
|
@@ -71,6 +71,6 @@ The citation in BibTeX is as follows.
|
|
| 71 |
year = {2026},
|
| 72 |
url = {https://petervartanian.xyz/aibd/},
|
| 73 |
doi = {10.57967/hf/11470},
|
| 74 |
-
note = {Edition 2026-11-12-
|
| 75 |
}
|
| 76 |
```
|
|
|
|
| 71 |
year = {2026},
|
| 72 |
url = {https://petervartanian.xyz/aibd/},
|
| 73 |
doi = {10.57967/hf/11470},
|
| 74 |
+
note = {Edition 2026-11-12-e97c02573b96}
|
| 75 |
}
|
| 76 |
```
|
adyt.csv
CHANGED
|
@@ -16,7 +16,7 @@ AIBD-0328,2026-04-05,event,day,reported,"A CIA tool called “Ghost Murmur”, p
|
|
| 16 |
AIBD-0302,2026-02-22,event,day,reported,"After Apple Watch Crash Detection calls 911 for a 19-year-old trapped upside down in her pickup in the Little Miami River, Clermont County crews pull her out with bruises only","On 22 February 2026 near Batavia, Ohio, a 19-year-old's truck flipped into the Little Miami River. Clermont County officials say that her Apple Watch Crash Detection called 911. Crews freed her.",Apple Watch Crash Detection calls 911 after a truck flips into a river,,,reported,Independently reported,no comparison,very low,protective response,false,Apple Watch Crash Detection,,Apple,United States,,,"Batavia Township, Clermont County, Ohio, United States",United States,Emergency response / road traffic injury,Consumer wearables,Safety and rescue,Prediction from data,Drivers and travelers,3,High income,physical_health,,5,1,5,,4,1,,personal; organizational,5,1,"the same night crash into a river with no automated call, which the driver believed would have left nobody aware of the crash",The driver says that she did not think that anybody could hear her or knew what had happened.,"A dispatcher stayed on the line, and crews pulled her out with bruises only.",assumed,persisted,2026-03-11,shown,https://www.apple.com/newsroom/2022/09/apple-reveals-apple-watch-series-8-and-the-new-apple-watch-se/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""ABC News, citing Clermont County officials, and WLWT say that the Apple Watch's Crash Detection placed the 911 call. On its support page, Apple says that the feature detects severe crashes from sensor data on the device and calls automatically after a countdown if the wearer does not respond. The driver also told WKRC that a Life360 crash-detection service on her mother's phone called police and family. The watch may not have been the only alert. Only news reports are available, and no dispatch log or agency statement is on record.""}, {""label"": ""Comparison"", ""value"": ""the same night crash into a river with no automated call, which the driver believed would have left nobody aware of the crash"", ""url"": ""https://abcnews.com/GMA/News/19-year-pulled-sinking-car-details-rescue-responders/story?id=130747244""}, {""label"": ""Before"", ""value"": ""The driver says that she did not think that anybody could hear her or knew what had happened.""}, {""label"": ""After"", ""value"": ""A dispatcher stayed on the line, and crews pulled her out with bruises only.""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 17 |
AIBD-0580,2025-06-10,event,day,reported,Lynxight alerts lifeguards at Inglewood Pool to an unconscious swimmer who recovers fully after rescue,"Lynxight, a camera system at a Perth pool, alerted a lifeguard's smartwatch to a man who had sunk unconscious. Lifeguards rescued him with help from a swimmer. He recovered fully.","Lynxight alerts a Perth lifeguard to a sinking swimmer, who is rescued",,"On 10 June 2025, a man suffered a medical episode and sank in the outdoor pool at Stirling Leisure - Inglewood in Perth. Lynxight is a camera system built to detect swimmers in distress. It sent a lifeguard a smartwatch alert with the swimmer's location and images. Another swimmer noticed the emergency on his own and lifted the man to the surface.[1] The alerted lifeguard and other lifeguards brought him out of the water. The City of Stirling reports that he was conscious when paramedics took him away. It also reports that he later recovered fully.[2] Help came by two routes. The alert reached a lifeguard who had not seen the man sink.[3] A fellow swimmer noticed him independently. The sources do not show whether he would have survived without the alert.",reported,Independently reported,no comparison,very low,harm prevented,false,Lynxight AI pool monitoring with CCTV analysis and smartwatch alerts,,Lynxight,Israel,City of Stirling,Australia,"Stirling Leisure - Inglewood, Perth, Western Australia, Australia",Australia,Aquatic safety and rescue,Public swimming pools,Safety and rescue,Images and video,Other workers; The public,3,High income,physical_health,Hazard reduction,5,1,5,,4,1,,personal,4,1,the lifeguard's own visual watch of the pool without the alert,the lifeguard on patrol did not see the man go under,"the smartwatch alert gave the lifeguard the swimmer's location with images, and the city reports that it improved his response time",assumed,persisted,2025-12-31,shown,https://www.lynxight.com/legal/privacy-policy-pools-he,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Lynxight detected distress and alerted the lifeguard, who had not seen the man sink. Human rescuers brought him to safety. Whether he would have survived without the alert is unknown.""}, {""label"": ""Comparison"", ""value"": ""the lifeguard's own visual watch of the pool without the alert"", ""url"": ""https://www.stirling.wa.gov.au/city-and-council/news/power-of-anti-drowning-technology-on-display-at-stirling-leisure-%E2%80%93-inglewood-in-wa%E2%80%99s-first-ai-assist""}, {""label"": ""Before"", ""value"": ""the lifeguard on patrol did not see the man go under""}, {""label"": ""After"", ""value"": ""the smartwatch alert gave the lifeguard the swimmer's location with images, and the city reports that it improved his response time""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 18 |
AIBD-0582,2024-02-02,event,day,reported,"After AI detects her brain hemorrhage, a nurse returns to nursing following aneurysm treatment","Software, apparently Viz.ai, detected a nurse’s brain bleed. It alerted her doctors. After treatment, the nurse returned to full-time nursing.",A nurse returns to full-time work after software detects her brain bleed,,"A 57-year-old nurse suffered a ruptured brain aneurysm. Her symptoms began on 2 February 2024.[1] According to her surgeon, software at Baptist Health's Marcus Neuroscience Institute flagged the bleeding on her scan and alerted clinicians.[2] The clinicians sealed the aneurysm with coils delivered through a catheter. After about three weeks in the hospital, she recovered and returned to full-time nursing.[3] In a later interview, the surgeon named Viz.ai. The surgeon also described an unnamed intensive care nurse whose circumstances match hers. The identification of the software therefore remains an inference.[4] No source isolates what the alert changed for her survival.",reported,Independently reported,no comparison,very low,harm prevented,false,Viz.ai stroke detection and care coordination (attribution inferred),,Viz.ai (inferred),United States,"Baptist Health, Marcus Neuroscience Institute",United States,"Boca Raton, Florida, United States",United States,Stroke care,Emergency patients with brain hemorrhage,Health,Images and video,Patients; Health workers,3,High income,physical_health,Medical diagnosis,5,1,5,,3,1,,personal,5,0,emergency physicians and radiologists reading the same scan,the bleeding had not yet been detected by emergency room doctors or radiologists,the AI picked up the bleeding before emergency room doctors or radiologists had detected it,assumed,persisted,2025-04-29,shown,https://www.accessdata.fda.gov/cdrh_docs/pdf19/K193658.pdf,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""AI flagged the bleeding and notified the stroke team, according to her treating surgeon. Clinicians treated the aneurysm. The identification as Viz.ai is inferred from a separate interview with the surgeon. No study in the sources isolates the software's effect on survival.""}, {""label"": ""Comparison"", ""value"": ""emergency physicians and radiologists reading the same scan"", ""url"": ""https://cbs12.com/news/local/from-caregiver-to-cared-for-how-an-icu-nurses-training-saved-her-own-life""}, {""label"": ""Before"", ""value"": ""the bleeding had not yet been detected by emergency room doctors or radiologists""}, {""label"": ""After"", ""value"": ""the AI picked up the bleeding before emergency room doctors or radiologists had detected it""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 19 |
-
AIBD-0296,2016-11-01,event,month,reported,"A seizure alert from an Empatica Embrace reaches the mother of a 26-year-old teacher, who calls paramedics after a seizure left her daughter not breathing (Fulton, Missouri)","An Embrace wristband detected a seizure that stopped a 26-year-old teacher's breathing. It alerted her mother, who called paramedics. IEEE Spectrum reports that she survived.","Embrace wristband alerts a mother, who calls paramedics",,"In November 2016 a 26-year-old teacher from Fulton, Missouri, had an epileptic seizure. It left her unresponsive and not breathing.[1] Her Embrace wristband detected the seizure and alerted her mother, who called paramedics. The teacher survived. Her rescue rests on a single IEEE Spectrum report. The U.S. Food and Drug Administration cleared the device on 26 January 2018, after the event.[2] Later evidence comes from hospital units where patients with epilepsy are monitored. Researchers in Montreal published a study of the device in their own unit on 7 January 2026. They found that the Embrace2 detected all 16 convulsive seizures that occurred while it was worn and working. It raised 29 false alarms in 373 days of monitoring. It detected none of 510 focal seizures without convulsions, a type that the device is not cleared to detect.[3]",reported,Independently reported,,very low,harm prevented,false,"Empatica Embrace wrist-worn seizure detection (electrodermal activity and accelerometer, on-board algorithm) with the Alert smartphone app",Embrace (the article does not state the generation); echoes concern Embrace2,Empatica,United States,,,"Fulton, Missouri, United States",United States,Healthcare (epilepsy; seizure monitoring),Prescription wearable seizure monitor used at home,Health,Prediction from data,Patients,3,High income,physical_health,,5,1,5,,5,1,,personal,5,3,"seizures recorded during more than 6,500 hours of monitoring of 135 patients (device test reported by IEEE Spectrum)",,"the device correctly identified nearly every single seizure, with a false alarm rate below that of the seizure frequency for most individuals",measured,persisted,2026-01-07,shown,https://pmc.ncbi.nlm.nih.gov/articles/PMC8418082/,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""According to IEEE Spectrum, the Embrace detected the seizure and its app alerted McGrath's mother, who made the emergency call. The device did not call for help itself. IEEE Spectrum calls the algorithm proprietary. In its summary, the FDA describes an on-board algorithm validated against video-EEG, and in their own study of 2021 Empatica's authors call it a predefined machine-learning algorithm. The McGrath event is known only from this single news item.""}, {""label"": ""Comparison"", ""value"": ""seizures recorded during more than 6,500 hours of monitoring of 135 patients (device test reported by IEEE Spectrum)"", ""url"": ""https://spectrum.ieee.org/this-seizuredetecting-smartwatch-could-save-your-life""}, {""label"": ""After"", ""value"": ""the device correctly identified nearly every single seizure, with a false alarm rate below that of the seizure frequency for most individuals""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""measured""}"
|
| 20 |
AIBD-1234,2026-11-05,source publication,day,reported,"Diario de Mallorca reports that QP Prostate marked a tiny lesion on a 58-year-old man's MRI at Clínica Girona, a biopsy confirmed cancer, and after surgery he says he is free of disease","Diario de Mallorca reports that QP Prostate marked a tiny lesion on a 58-year-old man's MRI at Clínica Girona. A biopsy confirmed cancer, and after surgery he says he is free of disease.",*QP Prostate* flags a tiny tumor that a biopsy confirms in Girona,,,reported,Independently reported,no comparison,very low,harm prevented,false,QP Prostate,,Quibim,Spain,Clínica Girona (radiology department),Spain,"Girona, Spain",Spain,Healthcare (radiology),,Health,Images and video,Patients; Health workers,3,High income,physical_health,Medical diagnosis,4,1,,,3,1,,personal,1,0,,,,,unknown,,not shown,,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""The tool marked the lesion and the radiologist advised the biopsy. No source compares the reading with and without the tool.""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""undocumented""}"
|
| 21 |
AIBD-0209,2026-06-09,event,day,reported,Saronic's Corsair sea drone carries two Apache crew members off Oman to a helicopter hoist,A remotely piloted Saronic Corsair sea drone carried two Apache crew off Oman to a helicopter hoist. Central Command reported that both were rescued within approximately two hours.,Saronic's Corsair sea drone picks up two Apache crew members off Oman,both were rescued within approximately two hours,,reported,Independently reported,no comparison,very low,harm prevented,true,Saronic Corsair unmanned (autonomous) surface vessel operated by U.S. 5th Fleet's Task Force 59,Corsair; navigation-software version not reported,Saronic Technologies (Texas),United States,"U.S. Navy, 5th Fleet Task Force 59 (U.S. Naval Forces Central Command)",United States,"Off the coast of Oman, near the Strait of Hormuz",Oman,Military search and rescue / personnel recovery at sea,"Government, security, and defense (naval unmanned surface vessels)",Defense,Machines that act,Public servants,3,High income,physical_health,,4,2,5,,3,1,,personal; organizational,5,0,"a rescue by a crewed platform, such as the crewed helicopters and planes used to recover the crew of an F-15 shot down over Iran in April 2026",crewed rescue craft put more personnel at risk; during the April rescue an A-10 was hit by anti-aircraft fire,"an unmanned vessel reached and carried the two soldiers, and a helicopter completed the recovery in safer waters",assumed,persisted,2026-09-18,not shown,https://www.al-monitor.com/originals/2026/06/drone-rescue-highlights-us-navys-autonomous-push,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Hawkins identified an unmanned Corsair as the recovery vessel. Al-Monitor reported that military officials told U.S. media that an operator piloted the vessel from afar. Stars and Stripes reported that the Corsair was guided to the site, located the soldiers and carried them to safer waters. Neither the Navy nor Saronic has said which functions ran on their own or what the vessel's sensors contributed. A crewed helicopter completed the recovery.""}, {""label"": ""Comparison"", ""value"": ""a rescue by a crewed platform, such as the crewed helicopters and planes used to recover the crew of an F-15 shot down over Iran in April 2026"", ""url"": ""https://www.stripes.com/branches/navy/2026-06-10/sea-drones-us-navy-deployments-21929112.html""}, {""label"": ""Before"", ""value"": ""crewed rescue craft put more personnel at risk; during the April rescue an A-10 was hit by anti-aircraft fire""}, {""label"": ""After"", ""value"": ""an unmanned vessel reached and carried the two soldiers, and a helicopter completed the recovery in safer waters""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 22 |
AIBD-0010,2026-06-07,event,day,reported,"A suicide alert from Meta, routed to Uttar Pradesh Police, brings officers to a 25-year-old in Meerut within eight minutes","After Meta alerted police to a 25-year-old man's video on Instagram, officers in Meerut reached his home within eight minutes. His condition improved after treatment in hospital.",Meta's suicide alert brings Meerut police to a man within 8 minutes,a Sardhana team reached his home within eight minutes,"Police said that Meta sent a suicide alert at 6:53 pm on 7 June 2026 about a 25-year-old Meerut man's Instagram video. By their account, a Sardhana police team reached him within eight minutes. The team took him to a hospital, where his condition improved.[1] The arrangement between Meta and Uttar Pradesh Police has run since an agreement signed in March 2022.[2] The police credit the alerts with 3,011 lives saved statewide from 1 January 2023 to 31 May 2026.[3] For the period to 10 October 2026 the figure is more than 4,000. Those counts come from the police. No outside audit of them is known. Police have not said whether Meta's classifier flagged the post or a user reported it. Meta publishes no false-positive data for India. That gap leaves open how often an alert sends officers to someone who was never at risk.[4]",reported,Independently reported,no comparison,very low,harm prevented,false,Meta proactive self-harm detection on Instagram with alert routing to the Uttar Pradesh Police Social Media Center,,Meta Platforms,United States,Meta Platforms with the Uttar Pradesh Police Social Media Center (agreement signed March 2022),United States,"Meerut, India",India,Suicide prevention / online safety,Social media platforms and police,Safety and rescue,Prediction from data,Patients,3,Lower middle income,physical_health,,4,1,5,,5,1,,personal,7,7,,,,,persisted,2026-10-22,shown,https://about.fb.com/news/2018/09/inside-feed-suicide-prevention-and-ai/,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Police officers said that Meta sent the alert after the post. They did not say whether Meta's classifier or a user's report flagged it. Meta and NPR describe the process as detection by machine learning, followed by human review before anyone contacts the authorities. The AI's part here rests on that general description.""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""undocumented""}"
|
|
@@ -93,9 +93,9 @@ AIBD-0356,2019-07-01,event,month,reported,"Rainforest Connection's Guardian, an
|
|
| 93 |
AIBD-0246,2026-06-24,event,day,reported,Odakyu Electric Railway starts an AI crossing image-analysis system at four Odawara Line crossings that signals approaching trains to stop when people are left inside,Odakyu's AI crossing image-analysis system reads existing crossing cameras. It signals approaching trains to stop when someone is left inside. Trains made 17 emergency stops in two months.,Odakyu trains stop when software detects a person left in a crossing,trains made 17 emergency stops on detected anomalies in the system's first two months,,reported,Independently reported,no comparison,very low,protective response,false,AI踏切画像解析システム (AI crossing image-analysis system) using existing crossing cameras linked to signaling equipment,,"Meitetsu EI Engineer, Toyota Systems and Toho Electric Industries",Japan,Odakyu Electric Railway,Japan,"Odakyu Odawara Line, Japan",Japan,Railway safety,Level-crossing obstacle detection,Safety and rescue,Images and video,Drivers and travelers,11,High income,physical_health,Hazard reduction,,,3,,4,5,,societal; organizational; personal,4,4,Detection by the train driver's own sight or by someone pressing the crossing's emergency button,People left inside a crossing are noticed by the driver's sight or the emergency button,The AI sends a stop signal to approaching trains; 17 emergency stops in the first two months,assumed,persisted,2026-09-19,shown,https://www.toyotasystems.com/news/upload/2026/06/9230e25b62c7992c0aa938759e7a7fd9589f2be2.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The AI detects people left inside the crossing and triggers the stop signal, and the crew also sees a warning lamp. Yomiuri reports 17 stops, but it does not say how many came from AI detections, how many were real hazards or how often the system missed a person.""}, {""label"": ""Comparison"", ""value"": ""Detection by the train driver's own sight or by someone pressing the crossing's emergency button"", ""url"": ""https://www.yomiuri.co.jp/national/20260919-GYT1T00229/2/""}, {""label"": ""Before"", ""value"": ""People left inside a crossing are noticed by the driver's sight or the emergency button""}, {""label"": ""After"", ""value"": ""The AI sends a stop signal to approaching trains; 17 emergency stops in the first two months""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 94 |
AIBD-0922,2026-06-01,source publication,month,reported,"UNDP's evaluation office reports that at three Zimbabwean irrigation schemes, drones with AI monitored crops, enabling earlier responses to threats",UNDP's evaluation office reported that drones with AI monitored crop health at three Zimbabwean irrigation schemes. This sped up responses to threats of crop loss. The responses were more precise.,Drones with AI watch crop health at three Zimbabwean irrigation schemes,,,reported,Independently reported,no comparison,very low,protective response,false,AI and machine-learning crop health monitoring with drone technology (product not named),,Not named in the evaluation,,"UNDP Zimbabwe, through the resilience-building project funded by the Green Climate Fund and run with the Government of Zimbabwe",Zimbabwe,"Midlo (Matabeleland South), Pikinini/Jawanda and Bindamombe (Masvingo) irrigation schemes, Zimbabwe",Zimbabwe,Agriculture / crop protection,,Agriculture and food,Images and video,Farmers,2,Lower middle income,environmental_impact,Environmental impacts,,,,4,3,,,personal,2,0,,,,,unknown,,shown,https://erc.undp.org/evaluation/documents/download/25595,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The office credits artificial intelligence, drone technology and machine-learning models together and does not separate their parts. It names no product or model and gives no figure for the earlier and more precise responses. The project's interim evaluation of September 2024 does not mention drones or artificial intelligence.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 95 |
AIBD-0245,2026-04-24,event,day,reported,"After a Tesla on Autopilot stops in the middle lane of I-75 with its drunk driver asleep, police arrest her for driving under the influence",Troopers said that the eye detection of a Tesla on Autopilot halted the car on I-75 after its intoxicated driver fell asleep. No vehicle data have been published to confirm it.,Tesla halts on I-75 when its drunk driver falls asleep at the wheel,,,reported,Independently reported,no comparison,very low,protective response,false,"Tesla Autopilot, according to the reported account; the vehicle model and active configuration are unverified.",,Tesla,United States,Private owner,,"Sarasota, United States",United States,Road vehicle crash avoidance,Consumer driver assistance / driver monitoring,Safety and rescue,Machines that act; Images and video,Drivers and travelers,3,High income,physical_health,,,1,1,,5,1,,personal; societal; organizational,4,2,,,,,unknown,,shown,https://en.wikipedia.org/wiki/Tesla_Autopilot,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Florida Highway Patrol troopers attribute the stop to the car's eye detection. WWSB reports that the car halted in the middle lane when the driver stopped responding to its alerts. The driver's decision to rely on Autopilot while drunk created the danger. No vehicle data were published to confirm the sequence.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 96 |
-
AIBD-0325,2025-05-24,source publication,day,unclear,A hospital's account conflicts with a relative’s claim that ChatGPT helped save a teenager from poisoning,A relative credited ChatGPT's advice to drink paraffin oil with saving a teenager from a grain pill. The medical dean at Assiut University said that the teenager showed no signs of poisoning.,Hospital disputes that ChatGPT helped save a teenager from poisoning,,,reported,Independently reported,,
|
| 97 |
-
AIBD-0326,2025-05-07,event,day,unclear,"Rafael's X-Guard towed decoy, labeled “AI-powered”, may have diverted Pakistani missiles from Indian Rafales in Operation Sindoor (7–10 May 2025). India's defense chief reports Indian aircraft losses. So does France's air force chief","Commentators credit Rafael's X-Guard towed decoy, labeled “AI-powered”, with diverting Pakistani missiles from Indian Rafales. No official has confirmed its use. India admitted jet losses.",X-Guard decoy may have diverted Pakistani missiles from Indian jets,,,reported,Independently reported,,
|
| 98 |
-
AIBD-2210,2024-06-11,source publication,day,intended,Equinoct: Integrating Community Data with Flood Forecasting for a Climate-Resilient Future,"Equinoct, an Indian company, joins rainfall and river readings that residents collect with AI to forecast floods in two river basins of the Western Ghats. No forecast result is reported.",*Equinoct* builds flood forecasts from villagers' rain gauges in Kerala,,,reported,Independently reported,,
|
| 99 |
AIBD-0371,2024-05-01,event,month,reported,"Within a week of tying up with Meta to route flags of self-harm from Facebook and Instagram to Abhay Command Centre, Kota police say that they prevented a possible suicide by a Jhunjhunu student","Kota police receive Meta's red-flag tags for posts about self-harm on Facebook and Instagram. Within a week, police said that they had prevented a possible suicide by a Jhunjhunu student.",Kota police act on a self-harm flag from Meta within a week of tie-up,,,reported,Independently reported,no comparison,very low,protective response,false,"Meta suicide and self-injury detection on Facebook and Instagram, with red-flag tags shared with Kota police and Abhay Command Centre",,Meta Platforms,United States,"Kota City Police (Rajasthan Police), Abhay Command Centre",India,"Kota, India",India,Suicide prevention / online safety,Social media platforms and police,Safety and rescue,Prediction from data,Students; Public servants,3,Lower middle income,emotional_health; physical_health,Emotional analysis/intelligence,,,,,5,,,personal,4,2,,,,,unknown,,shown,https://about.fb.com/news/2018/09/inside-feed-suicide-prevention-and-ai/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""PTI reports that the chain starts with a red-flag tag on Meta's system when a student posts suicide or self-harm content. Account information is then shared with police. PTI does not say whether Meta's classifier or a user report raised the flag in the Jhunjhunu case, or what happened on arrival. In its own description (2018, updated 2021), Meta says that machine-learning detection is followed by human review before authorities are contacted. The intervention itself was police work. PTI also notes that no red-flag tags had been reported in Kota city since the launch.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 100 |
AIBD-0113,2022-05-26,event,day,reported,"A threat alert from Meta, relayed through the FBI, brings Paraguayan police with psychologists to a bullied 17-year-old who praised the Uvalde shooter","Meta flagged a post in which a Paraguayan teenager praised the Uvalde shooter. After the FBI relayed the alert, police visited him with psychologists.","After Meta flags a post, police visit a Paraguayan teen",,,reported,Independently reported,no comparison,very low,protective response,false,Meta (Facebook/Instagram) violence-threat detection and law-enforcement referral process,,Meta Platforms,United States,Meta Platforms,United States,Paraguay,Paraguay,Online safety and threat referral to law enforcement,"Media, social platforms, and marketing",Safety and rescue,Prediction from data,Children,16,Upper middle income,physical_health; emotional_health,Crime reduction or prevention,,1,1,,4,1,,personal; societal,7,1,,,,,unknown,,shown,https://transparency.meta.com/enforcement/detecting-violations/how-enforcement-technology-works/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Officials said that the network detected the comment, classified it as dangerous and reported it. The report went to the FBI and to a 24/7 channel that prosecutors and the cybercrime police share. Alarcón also spoke of corroborating “datos de analítica artificial” before the visit. Neither the officials nor Meta says that the flag was automated. Meta says that its technology detects most violating content before anyone reports it, and that people review its referrals to law enforcement. Paraguayan police, prosecutors and the child-protection office identified the user, visited his home and recommended psychological support.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 101 |
AIBD-0111,2022-01-05,event,day,reported,Google's Android Earthquake Alerts reached phones in Alanya seconds before a quake of magnitude 5.1 in the Mediterranean on 5 January 2022,"Google's Android Earthquake Alerts System sent Android users near Alanya, Turkey, a warning seconds before a magnitude 5.1 offshore quake on 5 January 2022.",Android alerts reach phones in Alanya seconds before an offshore quake,Hürriyet put the lead at about 10 seconds,,reported,Independently reported,no comparison,very low,protective response,false,Android Earthquake Alerts System (crowdsourced phone-accelerometer detection; alerts for magnitude 4.5 and above),,Google,United States,Google,United States,"Alanya, Turkey",Turkey,Disaster risk reduction / earthquake early warning,Consumer smartphones (Android),Humanitarian,Prediction from data,People at risk of disaster,11,Upper middle income,environmental_impact,Natural disaster prediction,,,1,,4,1,,personal,16,2,"Feeling the shaking with no advance notice, as before the system reached Turkey",No warning before the shaking,"A phone alert about 10 seconds before the shaking, by Hürriyet's account; other outlets gave up to one minute",assumed,challenged,2023-07-27,not shown,https://blog.google/products/android/earthquake-detection-and-alerts/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Google's system produced the alert from start to finish. Its servers combined signals from the accelerometers of many phones to detect the quake and estimate its size. They then pushed alerts to phones where shaking was expected, ahead of the slower S-waves. Google does not use the words AI or machine learning when it describes this process. In a paper in Science in 2025, the authors describe it as algorithms that detect earthquakes with smartphones. Reporters give different lead times (seconds, about 10 seconds, about one minute), and Google promises seconds.""}, {""label"": ""Comparison"", ""value"": ""Feeling the shaking with no advance notice, as before the system reached Turkey"", ""url"": ""https://www.hurriyet.com.tr/gundem/erken-uyari-sistemi-tartismasi-depremin-olacagini-10-saniye-onceden-bilmek-guvenli-mi-41976168""}, {""label"": ""Before"", ""value"": ""No warning before the shaking""}, {""label"": ""After"", ""value"": ""A phone alert about 10 seconds before the shaking, by Hürriyet's account; other outlets gave up to one minute""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
|
@@ -134,7 +134,7 @@ AIBD-0082,2021-07-21,event,day,reported,"With ALICE, the CGU cuts R$2.38 million
|
|
| 134 |
AIBD-1725,2020-01-01,event,year,reported,Logically Intelligence,"Logically says that its AI service for tracking misinformation found 40,000 threats to election integrity for one US state in 2020 and raised violation detection by over 200% for one Indian state.","*Logically Intelligence* flags 40,000 election threats for a US state",200%,,claimed,The makers' claim,no comparison,very low,protective response,false,Logically Intelligence,,Logically,United Kingdom,Logically,United Kingdom,United States and India,United States | India,Government monitoring of online misinformation,,Public sector,Text and language,Public servants,16,Lower middle income; High income,social_improvement,Fact checking,2,,,,3,,,organizational,1,0,,,,,unknown,,shown,https://ircai.org/top100/entry/logically-intelligence/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Machine learning finds and groups suspect content and accounts. Analysts investigate, and clients decide whether to flag content to online services.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 135 |
AIBD-1748,2026-11-01,source access,month,reported,Real-Time Threat Detection,"Rainforest Connection says that its solar-powered Guardian devices, which use AI to find chainsaws, gunshots and engines in forest sound, have protected over 750,000 hectares of rainforest to date.","Rainforest Connection says that its devices guard 750,000 hectares",,,claimed,The makers' claim,no comparison,very low,protective response,false,Guardian (Rainforest Connection),,Rainforest Connection,United States,Rainforest Connection and local partners,,Rainforests in several countries (not listed on the page),,Forest conservation (acoustic detection of logging and poaching),,Environment,Speech and sound,,15,,environmental_impact,Environmental impacts,1,,,,3,,,societal,1,0,,,,,unknown,,shown,https://rfcx.org/guardian,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The system finds the sounds of threats and sends alerts. Local partners decide whether and how to respond.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 136 |
AIBD-0359,2026-09-01,event,month,reported,"Warning signs triggered by AI software that detects wildlife in thermal camera feeds slow drivers by an average of 3 mph on an elk corridor near Grundy, Virginia","In a Virginia pilot, Crosstek software detected wildlife in thermal camera feeds with 97 percent accuracy. Drivers slowed by an average of 3 mph at the warning signs that it activated.",Drivers slowed down when signs warned of wildlife ahead,an average speed reduction of 3 mph as vehicles approached activated signs,,claimed,The deployer's account,no method given,very low,protective response,false,"Animal Detection and Driver Warning System (ADDWS), using Crosstek thermal-camera AI detection",,Crosstek (cameras and detection software); system integration and evaluation by Virginia Tech Transportation Institute,United States,Virginia Department of Transportation with Virginia Tech Transportation Institute,United States,"Route 460 (Corridor Q) near Grundy, Buchanan County, Virginia, USA",United States,Road safety and wildlife-vehicle collision prevention,State transportation agency research pilot,Environment,Images and video,Drivers and travelers,11,High income,physical_health; animal_health,Hazard reduction,1,,4,,4,5,,personal; organizational; societal,4,2,a baseline that Virginia Tech does not specify,,an average speed reduction of 3 mph as vehicles approached signs that had been activated by wildlife detections,measured,unknown,,shown,https://news.vt.edu/articles/2026/10/vtti-elk-live-driver-warning-detection-virginia.html,"$80,771.49 per elk crash","Estimated comprehensive societal cost of an elk-vehicle crash in 2020 U.S. dollars, including economic loss and monetized pain and suffering. Context only: no AI-attributed savings or crashes prevented are established.",True,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""AI analysis of thermal camera feeds triggered the warning signs. Virginia Tech reports a 3 mph average slowdown as drivers approached activated signs, but it gives no result on collisions prevented. The results come from the evaluating institution's own report. In a separate interview with VDOT, WCYB covered the deployment and added no independent test of the measured results.""}, {""label"": ""Comparison"", ""value"": ""a baseline that Virginia Tech does not specify"", ""url"": ""https://news.vt.edu/articles/2026/10/vtti-elk-live-driver-warning-detection-virginia.html""}, {""label"": ""After"", ""value"": ""an average speed reduction of 3 mph as vehicles approached signs that had been activated by wildlife detections""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
| 137 |
-
AIBD-2190,2026-06-01,event,month,mixed,The night Venezuela shook: How Open Mapping Data Made the Emergency Response Faster,"The Humanitarian OpenStreetMap Team used its fAIr tool to rank areas and predict damage after Venezuela's earthquake of June 2026. In a test, mapping by hand was faster and more accurate.","*fAIr* ranked areas after Venezuela's quake, and a test favors hand work",,,claimed,The makers' claim,,very low,protective response,false,fAIr,,Humanitarian OpenStreetMap Team,,Humanitarian OpenStreetMap Team and its volunteer mappers,Venezuela,Venezuela (Caracas and La Guaira),Venezuela,Humanitarian mapping (building footprints and damage from satellite images),,Humanitarian,Images and video,Other workers,11,Lower middle income,operational_improvement; social_improvement,"Faster data management, decision-making",1,,,,2,,,organizational,2,0,Mapping by hand in the JOSM editor,86.08% of buildings mapped correctly by hand,78.22% of buildings mapped correctly with fAIr,measured,unknown,,shown,https://www.hotosm.org/en/news/the-night-venezuela-shook-how-open-mapping-data-made-the-emergency-response-faster/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models predict building footprints, density and damage from images. Volunteers correct and confirm every prediction before it is used.""}, {""label"": ""Comparison"", ""value"": ""Mapping by hand in the JOSM editor"", ""url"": ""https://media.ccc.de/v/sotm2025-osm-science-72051-user-testing-of-ai-assisted-mapping-tool-fair""}, {""label"": ""Before"", ""value"": ""86.08% of buildings mapped correctly by hand""}, {""label"": ""After"", ""value"": ""78.22% of buildings mapped correctly with fAIr""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
| 138 |
AIBD-0205,2026-04-01,source publication,month,reported,"Southern Water's sewer monitors flag a forming fatberg in Rowner, Gosport, so a crew clears it before any spill into the River Alver","Southern Water's machine-learning sewer monitors flagged a forming fatberg in Rowner, Gosport. A crew cleared it before any overflow into the River Alver.","Sensors flag a Gosport fatberg, which a crew clears before a spill",,,claimed,The deployer's account,no comparison,very low,protective response,false,"Southern Water's AI-enabled digital sewer system: a network of around 34,000 sewer level monitors in blockage hotspots with machine learning that learns normal sewer behavior and flags a forming blockage",,,,Southern Water,United Kingdom,"Gosport, United Kingdom",United Kingdom,Wastewater networks / pollution prevention,Regulated water and wastewater utility,Energy and infrastructure,Prediction from data,Other workers,6,High income,environmental_impact,Environmental impacts,1,1,1,,5,5,,societal; organizational,4,3,the previous reactive practice of attending blockages after a flooding or pollution event,turning up after the event to clean up and commiserate with devastated customers,spotting hundreds of potential blockages before it is too late,assumed,persisted,2026-10-08,shown,https://www.dailyecho.co.uk/news/26045031.fatberg-found-gosport-removed-causing-damage/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The level monitors detected an abnormal flow at Rowner, and the machine-learning model raised the alert to the control room. By the company's account, it distinguishes a forming blockage from daily rushes and rainfall. The jetting crew found and removed the fatberg. The articles rest on a Southern Water statement and one quoted manager. Reporters do not name the developer of the model and give no sensor data. No independent party confirmed that an overflow was imminent.""}, {""label"": ""Comparison"", ""value"": ""the previous reactive practice of attending blockages after a flooding or pollution event"", ""url"": ""https://www.dailyecho.co.uk/news/26045031.fatberg-found-gosport-removed-causing-damage/""}, {""label"": ""Before"", ""value"": ""turning up after the event to clean up and commiserate with devastated customers""}, {""label"": ""After"", ""value"": ""spotting hundreds of potential blockages before it is too late""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 139 |
AIBD-1670,2026-02-01,source publication,month,reported,AI-Driven Environmental Health Risk Prediction in Urban Morocco: The OCT7 SI Framework in Meknès,"A researcher at Moulay Ismail University says that OCT7 SI, which predicts periods of health risk in Meknès from weather, pollution and hospital admissions, supported earlier coordination in a pilot.",*OCT7 SI* warns Meknès health services of pollution and heat risk,,,claimed,The makers' claim,no comparison,very low,protective response,false,OCT7 SI (Operational Climate and Health Threat Intelligence System),,Moulay Ismail University of Meknès,Morocco,Moulay Ismail University of Meknès,Morocco,"Meknès, Morocco",Morocco,Health (surveillance of heat and air pollution risk),,Health,Prediction from data,,3; 13,Lower middle income,physical_health,Hazard reduction,1,4,,,3,,,organizational; personal,1,0,,,,,unknown,,shown,https://d19ob9sqegt2wc.cloudfront.net/stage/uploads/AI_Health_Compendium_2026_5604af298c.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models identify conditions that precede more hospital admissions. Health authorities decide on advisories and resources.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 140 |
AIBD-1673,2026-02-01,source publication,month,mixed,Spatiotemporal Probabilistic Encoder-fed Climate-aware Transfer-learning-ready Risk-forecasting Adversarial (SPECTRA) Architecture for Climate-informed Infectious Disease Forecasting,"Malaria No More's institute says that SPECTRA, a forecast model for climate-sensitive diseases, runs in national early warning systems in Indonesia and Bangladesh. No effect on outbreaks is shown.",*SPECTRA* forecasts feed disease warnings in Indonesia and Bangladesh,,,claimed,The makers' claim,no comparison,very low,protective response,false,SPECTRA,,"Institute for Health Modeling and Climate Solutions (Malaria No More), with the Indian Institute of Information Technology Vadodara",United States,Health authorities of Indonesia and Bangladesh,"Multiple: Indonesia, Bangladesh",Indonesia and Bangladesh,Indonesia | Bangladesh,Health (infectious disease early warning),,Health,Prediction from data,Public servants,3; 13,Lower middle income; Upper middle income,physical_health,Hazard reduction,1,,,,2,,,organizational; personal,1,0,,,,,unknown,,shown,https://d19ob9sqegt2wc.cloudfront.net/stage/uploads/AI_Health_Compendium_2026_5604af298c.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The model forecasts disease risk with uncertainty bounds. Public health teams decide on alerts and action.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
|
@@ -158,8 +158,8 @@ AIBD-1909,2021-01-01,event,year,reported,Factmata Narrative Monitoring,"Factmata
|
|
| 158 |
AIBD-1955,2021-01-01,source publication,year,reported,Wildfire Spread Prediction,"Fion, a U.S. startup, says that its deep learning maps wildfire risk, finds fire perimeters in satellite images and forecasts spread over 24 hours, for two large insurers and a U.S. state.",*Fion* forecasts wildfire spread for two insurers and a U.S. state,,,claimed,The makers' claim,no comparison,very low,protective response,false,Fion wildfire intelligence platform,,Fion Technologies Inc.,United States,Fion Technologies Inc.,United States,United States,United States,Disaster risk (wildfire spread forecasting),,Humanitarian,Images and video; Prediction from data,,13; 15,High income,knowledge_discovery; physical_health,Natural disaster prediction,1,2,,,3,,,organizational,1,0,,,,,unknown,,shown,https://ircai.org/top100/entry/wildfire-spread-prediction/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models map risk, find fire perimeters and forecast spread. Insurers and officials decide what to do.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 159 |
AIBD-1709,2019-01-01,event,year,reported,The ITIKI project combines AI with indigenous knowledge to predict droughts,"The GSMA reports that ITIKI, a drought prediction tool in South Africa, Kenya and Mozambique, sends forecasts to farmers by text message and claims 98% accuracy up to 19 months ahead.",*ITIKI* joins sensors and indigenous signs to warn farmers of drought,98%,,claimed,The makers' claim,no comparison,very low,protective response,false,ITIKI,,Central University of Technology (ITIKI project),South Africa,Central University of Technology (ITIKI project),South Africa,"South Africa, Kenya and Mozambique",South Africa | Kenya | Mozambique,Agriculture (drought forecasts for farmers),,Agriculture and food,Prediction from data,Farmers,2; 13,Low income; Lower middle income; Upper middle income,environmental_impact,Natural disaster prediction,1,,4,,3,,,personal,1,0,,,,,unknown,,not shown,,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The algorithms combine sensor readings with indigenous signs into forecasts. Farmers choose what to plant.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 160 |
AIBD-0329,2025-08-28,event,day,reported,"WINGS says, quoting the mayor, that wi.breathe cameras alert authorities to a Troizina wildfire",WINGS says that the AI fire detector on its wi.breathe cameras alerted authorities to a wildfire at Troizina on 28 August 2025. The municipality did not mention them in its statement that day.,wi.breathe cameras reportedly flag a Troizina wildfire,,,claimed,The makers' claim,no method given,very low,protective response,false,"wi.breathe Operational Intelligence Platform (WINGS ICT Solutions): hybrid optical-thermal cameras, a WINGS-trained AI smoke and fire detection model, cloud infrastructure and 5G/4G links with dashboard, email and SMS alerts",,"WINGS ICT Solutions, Athens, Greece",Greece,"Municipality of Troizinia-Methana (cameras at Vigla on Poros, Tsepeleni Aderon and Agios Georgios Methanon)",Greece,"Troizinia-Methana and Poros, Attica, Greece",Greece,Civil protection: wildfire early detection,Camera-based AI smoke and fire detection,Safety and rescue,Images and video,Public servants,13,High income,environmental_impact,Natural disaster prediction,,,1,,4,1,,societal; organizational,5,1,"the same fire without the camera alert, as the mayor describes it",,the fire was contained before it spread out of control,assumed,scaled,2026-05-12,shown,https://www.wings-ict-solutions.eu/wings-ict-solutions-delivers-wi-breathe-for-early-wildfire-detection-and-enhanced-environmental-intelligence-in-troizinia-methana-poros/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""WINGS reports that its system detected the fire through the Methana cameras and issued real-time alerts, and the mayor credits the system in WINGS's release. In its own statement on the day, the municipality credits more than 80 firefighters, six aircraft in action within 30 minutes, the police and volunteers. It does not mention the cameras. Neither WINGS nor the municipality has given the time of the AI alert or compared it with the first call to the fire service. The AI's share in the response is therefore unknown.""}, {""label"": ""Comparison"", ""value"": ""the same fire without the camera alert, as the mayor describes it"", ""url"": ""https://www.wings-ict-solutions.eu/early-wildfire-detection-in-trizina-by-wings-wi-breathe-innovation-in-action/""}, {""label"": ""After"", ""value"": ""the fire was contained before it spread out of control""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 161 |
-
AIBD-2192,2025-07-09,source publication,day,intended,Meet the winners for the 2025 AI for Good Impact Awards,SidLabs built Defensa+AI without charge for the Peruvian Society for Environmental Law. It is meant to alert communities and authorities to illegal logging and mining. No result is reported.,*Defensa+AI* is built to warn Peru's Amazon defenders of illegal logging,,,claimed,The makers' claim,,
|
| 162 |
-
AIBD-2219,2024-12-16,source publication,day,intended,Similie: Developing Open-Source Machine Learning Models to Transform Early Warning Systems and Forecast Climate Risks,"Similie runs flood warning sensors for over 300,000 people in Timor-Leste. With UNICEF's fund it built an open machine learning model to forecast rain. The model has no reported result.",*Similie* adds a learning model to Timor-Leste's flood warnings,,,claimed,The deployer's account,,
|
| 163 |
AIBD-2214,2024-08-01,event,month,tested,Turning Local Data into Climate Action: How Map & Rank Is Empowering Communities in Cameroon,"Map & Rank, a Cameroonian company, built a machine learning model that predicts floods and droughts for communities in the far north. The company reports an accuracy of 60%.",Map & Rank's flood model for northern Cameroon is right 60% of the time,an accuracy of 60%,,claimed,The makers' claim,no comparison,very low,protective response,false,Residat,,Map & Rank,Cameroon,Map & Rank,Cameroon,"Mayo-Danay, Far North Region, Cameroon",Cameroon,Climate risk (prediction of floods and droughts for communities),,Environment,Prediction from data,People at risk of disaster,13,Lower middle income,physical_health; environmental_impact,Natural disaster prediction,,,,,,,,societal,1,0,,,,,unknown,,shown,https://www.unicefventurefund.org/story/turning-local-data-climate-action-how-map-rank-empowering-communities-cameroon,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The model predicts floods and droughts for an area. Communities and local bodies would decide how to prepare.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 164 |
-
AIBD-2205,2024-06-17,source publication,day,intended,AEDES: A platform for dengue prediction using climate and health data for epidemic management,"Cirrolytix, a Philippine company, built AEDES to forecast dengue from climate, health and search data and to show likely hotspots from satellite images. No use by health offices is reported.",*AEDES* is built to forecast dengue hotspots in the Philippines,,,claimed,The makers' claim,,
|
| 165 |
AIBD-0036,2023-12-27,event,day,reported,"An automated classifier on a glider detects right whale calls, which prompt a voluntary NOAA Slow Zone east of Portland, Maine","The Low-Frequency Detection and Classification System on a glider detected right whales east of Portland, Maine, on 27 December 2023. NOAA declared a voluntary Slow Zone.",NOAA asks mariners to slow after glider software detects right whales,route around the area or transit at 10 knots or less,,claimed,The deployer's account,no comparison,very low,protective response,false,DMON instrument running the Low-Frequency Detection and Classification System (LFDCS) on a Slocum glider,,Woods Hole Oceanographic Institution (Mark Baumgartner),United States,Woods Hole Oceanographic Institution and NOAA Northeast Fisheries Science Center; NOAA Fisheries issues Slow Zones,United States,"Gulf of Maine east of Portland, United States",United States,Marine conservation / shipping,Endangered-species vessel-strike mitigation,Environment,Speech and sound,,14,High income,animal_health; environmental_impact,Environmental impacts,,,2,,5,1,,societal,6,2,aerial surveillance,,more slowdowns than aerial surveillance by a factor of two to five during fall and summer,measured,scaled,2025-01-21,shown,https://dcs.whoi.edu/lfdcs_manual/lfdcs_manual.html,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Automated classification produced the candidate detections. Analysts validated them on shore before NOAA acted. The resulting Slow Zone is voluntary.""}, {""label"": ""Comparison"", ""value"": ""aerial surveillance"", ""url"": ""https://tos.org/oceanography/article/glider-surveillance-for-near-real-time-detection-and-spatial-management-of-north-atlantic-right-whales""}, {""label"": ""After"", ""value"": ""more slowdowns than aerial surveillance by a factor of two to five during fall and summer""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
|
|
|
| 16 |
AIBD-0302,2026-02-22,event,day,reported,"After Apple Watch Crash Detection calls 911 for a 19-year-old trapped upside down in her pickup in the Little Miami River, Clermont County crews pull her out with bruises only","On 22 February 2026 near Batavia, Ohio, a 19-year-old's truck flipped into the Little Miami River. Clermont County officials say that her Apple Watch Crash Detection called 911. Crews freed her.",Apple Watch Crash Detection calls 911 after a truck flips into a river,,,reported,Independently reported,no comparison,very low,protective response,false,Apple Watch Crash Detection,,Apple,United States,,,"Batavia Township, Clermont County, Ohio, United States",United States,Emergency response / road traffic injury,Consumer wearables,Safety and rescue,Prediction from data,Drivers and travelers,3,High income,physical_health,,5,1,5,,4,1,,personal; organizational,5,1,"the same night crash into a river with no automated call, which the driver believed would have left nobody aware of the crash",The driver says that she did not think that anybody could hear her or knew what had happened.,"A dispatcher stayed on the line, and crews pulled her out with bruises only.",assumed,persisted,2026-03-11,shown,https://www.apple.com/newsroom/2022/09/apple-reveals-apple-watch-series-8-and-the-new-apple-watch-se/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""ABC News, citing Clermont County officials, and WLWT say that the Apple Watch's Crash Detection placed the 911 call. On its support page, Apple says that the feature detects severe crashes from sensor data on the device and calls automatically after a countdown if the wearer does not respond. The driver also told WKRC that a Life360 crash-detection service on her mother's phone called police and family. The watch may not have been the only alert. Only news reports are available, and no dispatch log or agency statement is on record.""}, {""label"": ""Comparison"", ""value"": ""the same night crash into a river with no automated call, which the driver believed would have left nobody aware of the crash"", ""url"": ""https://abcnews.com/GMA/News/19-year-pulled-sinking-car-details-rescue-responders/story?id=130747244""}, {""label"": ""Before"", ""value"": ""The driver says that she did not think that anybody could hear her or knew what had happened.""}, {""label"": ""After"", ""value"": ""A dispatcher stayed on the line, and crews pulled her out with bruises only.""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 17 |
AIBD-0580,2025-06-10,event,day,reported,Lynxight alerts lifeguards at Inglewood Pool to an unconscious swimmer who recovers fully after rescue,"Lynxight, a camera system at a Perth pool, alerted a lifeguard's smartwatch to a man who had sunk unconscious. Lifeguards rescued him with help from a swimmer. He recovered fully.","Lynxight alerts a Perth lifeguard to a sinking swimmer, who is rescued",,"On 10 June 2025, a man suffered a medical episode and sank in the outdoor pool at Stirling Leisure - Inglewood in Perth. Lynxight is a camera system built to detect swimmers in distress. It sent a lifeguard a smartwatch alert with the swimmer's location and images. Another swimmer noticed the emergency on his own and lifted the man to the surface.[1] The alerted lifeguard and other lifeguards brought him out of the water. The City of Stirling reports that he was conscious when paramedics took him away. It also reports that he later recovered fully.[2] Help came by two routes. The alert reached a lifeguard who had not seen the man sink.[3] A fellow swimmer noticed him independently. The sources do not show whether he would have survived without the alert.",reported,Independently reported,no comparison,very low,harm prevented,false,Lynxight AI pool monitoring with CCTV analysis and smartwatch alerts,,Lynxight,Israel,City of Stirling,Australia,"Stirling Leisure - Inglewood, Perth, Western Australia, Australia",Australia,Aquatic safety and rescue,Public swimming pools,Safety and rescue,Images and video,Other workers; The public,3,High income,physical_health,Hazard reduction,5,1,5,,4,1,,personal,4,1,the lifeguard's own visual watch of the pool without the alert,the lifeguard on patrol did not see the man go under,"the smartwatch alert gave the lifeguard the swimmer's location with images, and the city reports that it improved his response time",assumed,persisted,2025-12-31,shown,https://www.lynxight.com/legal/privacy-policy-pools-he,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Lynxight detected distress and alerted the lifeguard, who had not seen the man sink. Human rescuers brought him to safety. Whether he would have survived without the alert is unknown.""}, {""label"": ""Comparison"", ""value"": ""the lifeguard's own visual watch of the pool without the alert"", ""url"": ""https://www.stirling.wa.gov.au/city-and-council/news/power-of-anti-drowning-technology-on-display-at-stirling-leisure-%E2%80%93-inglewood-in-wa%E2%80%99s-first-ai-assist""}, {""label"": ""Before"", ""value"": ""the lifeguard on patrol did not see the man go under""}, {""label"": ""After"", ""value"": ""the smartwatch alert gave the lifeguard the swimmer's location with images, and the city reports that it improved his response time""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 18 |
AIBD-0582,2024-02-02,event,day,reported,"After AI detects her brain hemorrhage, a nurse returns to nursing following aneurysm treatment","Software, apparently Viz.ai, detected a nurse’s brain bleed. It alerted her doctors. After treatment, the nurse returned to full-time nursing.",A nurse returns to full-time work after software detects her brain bleed,,"A 57-year-old nurse suffered a ruptured brain aneurysm. Her symptoms began on 2 February 2024.[1] According to her surgeon, software at Baptist Health's Marcus Neuroscience Institute flagged the bleeding on her scan and alerted clinicians.[2] The clinicians sealed the aneurysm with coils delivered through a catheter. After about three weeks in the hospital, she recovered and returned to full-time nursing.[3] In a later interview, the surgeon named Viz.ai. The surgeon also described an unnamed intensive care nurse whose circumstances match hers. The identification of the software therefore remains an inference.[4] No source isolates what the alert changed for her survival.",reported,Independently reported,no comparison,very low,harm prevented,false,Viz.ai stroke detection and care coordination (attribution inferred),,Viz.ai (inferred),United States,"Baptist Health, Marcus Neuroscience Institute",United States,"Boca Raton, Florida, United States",United States,Stroke care,Emergency patients with brain hemorrhage,Health,Images and video,Patients; Health workers,3,High income,physical_health,Medical diagnosis,5,1,5,,3,1,,personal,5,0,emergency physicians and radiologists reading the same scan,the bleeding had not yet been detected by emergency room doctors or radiologists,the AI picked up the bleeding before emergency room doctors or radiologists had detected it,assumed,persisted,2025-04-29,shown,https://www.accessdata.fda.gov/cdrh_docs/pdf19/K193658.pdf,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""AI flagged the bleeding and notified the stroke team, according to her treating surgeon. Clinicians treated the aneurysm. The identification as Viz.ai is inferred from a separate interview with the surgeon. No study in the sources isolates the software's effect on survival.""}, {""label"": ""Comparison"", ""value"": ""emergency physicians and radiologists reading the same scan"", ""url"": ""https://cbs12.com/news/local/from-caregiver-to-cared-for-how-an-icu-nurses-training-saved-her-own-life""}, {""label"": ""Before"", ""value"": ""the bleeding had not yet been detected by emergency room doctors or radiologists""}, {""label"": ""After"", ""value"": ""the AI picked up the bleeding before emergency room doctors or radiologists had detected it""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 19 |
+
AIBD-0296,2016-11-01,event,month,reported,"A seizure alert from an Empatica Embrace reaches the mother of a 26-year-old teacher, who calls paramedics after a seizure left her daughter not breathing (Fulton, Missouri)","An Embrace wristband detected a seizure that stopped a 26-year-old teacher's breathing. It alerted her mother, who called paramedics. IEEE Spectrum reports that she survived.","Embrace wristband alerts a mother, who calls paramedics",,"In November 2016 a 26-year-old teacher from Fulton, Missouri, had an epileptic seizure. It left her unresponsive and not breathing.[1] Her Embrace wristband detected the seizure and alerted her mother, who called paramedics. The teacher survived. Her rescue rests on a single IEEE Spectrum report. The U.S. Food and Drug Administration cleared the device on 26 January 2018, after the event.[2] Later evidence comes from hospital units where patients with epilepsy are monitored. Researchers in Montreal published a study of the device in their own unit on 7 January 2026. They found that the Embrace2 detected all 16 convulsive seizures that occurred while it was worn and working. It raised 29 false alarms in 373 days of monitoring. It detected none of 510 focal seizures without convulsions, a type that the device is not cleared to detect.[3]",reported,Independently reported,no comparison,very low,harm prevented,false,"Empatica Embrace wrist-worn seizure detection (electrodermal activity and accelerometer, on-board algorithm) with the Alert smartphone app",Embrace (the article does not state the generation); echoes concern Embrace2,Empatica,United States,,,"Fulton, Missouri, United States",United States,Healthcare (epilepsy; seizure monitoring),Prescription wearable seizure monitor used at home,Health,Prediction from data,Patients,3,High income,physical_health,,5,1,5,,5,1,,personal,5,3,"seizures recorded during more than 6,500 hours of monitoring of 135 patients (device test reported by IEEE Spectrum)",,"the device correctly identified nearly every single seizure, with a false alarm rate below that of the seizure frequency for most individuals",measured,persisted,2026-01-07,shown,https://pmc.ncbi.nlm.nih.gov/articles/PMC8418082/,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""According to IEEE Spectrum, the Embrace detected the seizure and its app alerted McGrath's mother, who made the emergency call. The device did not call for help itself. IEEE Spectrum calls the algorithm proprietary. In its summary, the FDA describes an on-board algorithm validated against video-EEG, and in their own study of 2021 Empatica's authors call it a predefined machine-learning algorithm. The McGrath event is known only from this single news item.""}, {""label"": ""Comparison"", ""value"": ""seizures recorded during more than 6,500 hours of monitoring of 135 patients (device test reported by IEEE Spectrum)"", ""url"": ""https://spectrum.ieee.org/this-seizuredetecting-smartwatch-could-save-your-life""}, {""label"": ""After"", ""value"": ""the device correctly identified nearly every single seizure, with a false alarm rate below that of the seizure frequency for most individuals""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""measured""}"
|
| 20 |
AIBD-1234,2026-11-05,source publication,day,reported,"Diario de Mallorca reports that QP Prostate marked a tiny lesion on a 58-year-old man's MRI at Clínica Girona, a biopsy confirmed cancer, and after surgery he says he is free of disease","Diario de Mallorca reports that QP Prostate marked a tiny lesion on a 58-year-old man's MRI at Clínica Girona. A biopsy confirmed cancer, and after surgery he says he is free of disease.",*QP Prostate* flags a tiny tumor that a biopsy confirms in Girona,,,reported,Independently reported,no comparison,very low,harm prevented,false,QP Prostate,,Quibim,Spain,Clínica Girona (radiology department),Spain,"Girona, Spain",Spain,Healthcare (radiology),,Health,Images and video,Patients; Health workers,3,High income,physical_health,Medical diagnosis,4,1,,,3,1,,personal,1,0,,,,,unknown,,not shown,,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""The tool marked the lesion and the radiologist advised the biopsy. No source compares the reading with and without the tool.""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""undocumented""}"
|
| 21 |
AIBD-0209,2026-06-09,event,day,reported,Saronic's Corsair sea drone carries two Apache crew members off Oman to a helicopter hoist,A remotely piloted Saronic Corsair sea drone carried two Apache crew off Oman to a helicopter hoist. Central Command reported that both were rescued within approximately two hours.,Saronic's Corsair sea drone picks up two Apache crew members off Oman,both were rescued within approximately two hours,,reported,Independently reported,no comparison,very low,harm prevented,true,Saronic Corsair unmanned (autonomous) surface vessel operated by U.S. 5th Fleet's Task Force 59,Corsair; navigation-software version not reported,Saronic Technologies (Texas),United States,"U.S. Navy, 5th Fleet Task Force 59 (U.S. Naval Forces Central Command)",United States,"Off the coast of Oman, near the Strait of Hormuz",Oman,Military search and rescue / personnel recovery at sea,"Government, security, and defense (naval unmanned surface vessels)",Defense,Machines that act,Public servants,3,High income,physical_health,,4,2,5,,3,1,,personal; organizational,5,0,"a rescue by a crewed platform, such as the crewed helicopters and planes used to recover the crew of an F-15 shot down over Iran in April 2026",crewed rescue craft put more personnel at risk; during the April rescue an A-10 was hit by anti-aircraft fire,"an unmanned vessel reached and carried the two soldiers, and a helicopter completed the recovery in safer waters",assumed,persisted,2026-09-18,not shown,https://www.al-monitor.com/originals/2026/06/drone-rescue-highlights-us-navys-autonomous-push,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Hawkins identified an unmanned Corsair as the recovery vessel. Al-Monitor reported that military officials told U.S. media that an operator piloted the vessel from afar. Stars and Stripes reported that the Corsair was guided to the site, located the soldiers and carried them to safer waters. Neither the Navy nor Saronic has said which functions ran on their own or what the vessel's sensors contributed. A crewed helicopter completed the recovery.""}, {""label"": ""Comparison"", ""value"": ""a rescue by a crewed platform, such as the crewed helicopters and planes used to recover the crew of an F-15 shot down over Iran in April 2026"", ""url"": ""https://www.stripes.com/branches/navy/2026-06-10/sea-drones-us-navy-deployments-21929112.html""}, {""label"": ""Before"", ""value"": ""crewed rescue craft put more personnel at risk; during the April rescue an A-10 was hit by anti-aircraft fire""}, {""label"": ""After"", ""value"": ""an unmanned vessel reached and carried the two soldiers, and a helicopter completed the recovery in safer waters""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""assumed""}"
|
| 22 |
AIBD-0010,2026-06-07,event,day,reported,"A suicide alert from Meta, routed to Uttar Pradesh Police, brings officers to a 25-year-old in Meerut within eight minutes","After Meta alerted police to a 25-year-old man's video on Instagram, officers in Meerut reached his home within eight minutes. His condition improved after treatment in hospital.",Meta's suicide alert brings Meerut police to a man within 8 minutes,a Sardhana team reached his home within eight minutes,"Police said that Meta sent a suicide alert at 6:53 pm on 7 June 2026 about a 25-year-old Meerut man's Instagram video. By their account, a Sardhana police team reached him within eight minutes. The team took him to a hospital, where his condition improved.[1] The arrangement between Meta and Uttar Pradesh Police has run since an agreement signed in March 2022.[2] The police credit the alerts with 3,011 lives saved statewide from 1 January 2023 to 31 May 2026.[3] For the period to 10 October 2026 the figure is more than 4,000. Those counts come from the police. No outside audit of them is known. Police have not said whether Meta's classifier flagged the post or a user reported it. Meta publishes no false-positive data for India. That gap leaves open how often an alert sends officers to someone who was never at risk.[4]",reported,Independently reported,no comparison,very low,harm prevented,false,Meta proactive self-harm detection on Instagram with alert routing to the Uttar Pradesh Police Social Media Center,,Meta Platforms,United States,Meta Platforms with the Uttar Pradesh Police Social Media Center (agreement signed March 2022),United States,"Meerut, India",India,Suicide prevention / online safety,Social media platforms and police,Safety and rescue,Prediction from data,Patients,3,Lower middle income,physical_health,,4,1,5,,5,1,,personal,7,7,,,,,persisted,2026-10-22,shown,https://about.fb.com/news/2018/09/inside-feed-suicide-prevention-and-ai/,,,,[],Adyt,Specific harm prevented.,"[{""label"": ""Contribution of the AI"", ""value"": ""Police officers said that Meta sent the alert after the post. They did not say whether Meta's classifier or a user's report flagged it. Meta and NPR describe the process as detection by machine learning, followed by human review before anyone contacts the authorities. The AI's part here rests on that general description.""}]","{""category"": ""Specific harm prevented"", ""comparison"": ""undocumented""}"
|
|
|
|
| 93 |
AIBD-0246,2026-06-24,event,day,reported,Odakyu Electric Railway starts an AI crossing image-analysis system at four Odawara Line crossings that signals approaching trains to stop when people are left inside,Odakyu's AI crossing image-analysis system reads existing crossing cameras. It signals approaching trains to stop when someone is left inside. Trains made 17 emergency stops in two months.,Odakyu trains stop when software detects a person left in a crossing,trains made 17 emergency stops on detected anomalies in the system's first two months,,reported,Independently reported,no comparison,very low,protective response,false,AI踏切画像解析システム (AI crossing image-analysis system) using existing crossing cameras linked to signaling equipment,,"Meitetsu EI Engineer, Toyota Systems and Toho Electric Industries",Japan,Odakyu Electric Railway,Japan,"Odakyu Odawara Line, Japan",Japan,Railway safety,Level-crossing obstacle detection,Safety and rescue,Images and video,Drivers and travelers,11,High income,physical_health,Hazard reduction,,,3,,4,5,,societal; organizational; personal,4,4,Detection by the train driver's own sight or by someone pressing the crossing's emergency button,People left inside a crossing are noticed by the driver's sight or the emergency button,The AI sends a stop signal to approaching trains; 17 emergency stops in the first two months,assumed,persisted,2026-09-19,shown,https://www.toyotasystems.com/news/upload/2026/06/9230e25b62c7992c0aa938759e7a7fd9589f2be2.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The AI detects people left inside the crossing and triggers the stop signal, and the crew also sees a warning lamp. Yomiuri reports 17 stops, but it does not say how many came from AI detections, how many were real hazards or how often the system missed a person.""}, {""label"": ""Comparison"", ""value"": ""Detection by the train driver's own sight or by someone pressing the crossing's emergency button"", ""url"": ""https://www.yomiuri.co.jp/national/20260919-GYT1T00229/2/""}, {""label"": ""Before"", ""value"": ""People left inside a crossing are noticed by the driver's sight or the emergency button""}, {""label"": ""After"", ""value"": ""The AI sends a stop signal to approaching trains; 17 emergency stops in the first two months""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 94 |
AIBD-0922,2026-06-01,source publication,month,reported,"UNDP's evaluation office reports that at three Zimbabwean irrigation schemes, drones with AI monitored crops, enabling earlier responses to threats",UNDP's evaluation office reported that drones with AI monitored crop health at three Zimbabwean irrigation schemes. This sped up responses to threats of crop loss. The responses were more precise.,Drones with AI watch crop health at three Zimbabwean irrigation schemes,,,reported,Independently reported,no comparison,very low,protective response,false,AI and machine-learning crop health monitoring with drone technology (product not named),,Not named in the evaluation,,"UNDP Zimbabwe, through the resilience-building project funded by the Green Climate Fund and run with the Government of Zimbabwe",Zimbabwe,"Midlo (Matabeleland South), Pikinini/Jawanda and Bindamombe (Masvingo) irrigation schemes, Zimbabwe",Zimbabwe,Agriculture / crop protection,,Agriculture and food,Images and video,Farmers,2,Lower middle income,environmental_impact,Environmental impacts,,,,4,3,,,personal,2,0,,,,,unknown,,shown,https://erc.undp.org/evaluation/documents/download/25595,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The office credits artificial intelligence, drone technology and machine-learning models together and does not separate their parts. It names no product or model and gives no figure for the earlier and more precise responses. The project's interim evaluation of September 2024 does not mention drones or artificial intelligence.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 95 |
AIBD-0245,2026-04-24,event,day,reported,"After a Tesla on Autopilot stops in the middle lane of I-75 with its drunk driver asleep, police arrest her for driving under the influence",Troopers said that the eye detection of a Tesla on Autopilot halted the car on I-75 after its intoxicated driver fell asleep. No vehicle data have been published to confirm it.,Tesla halts on I-75 when its drunk driver falls asleep at the wheel,,,reported,Independently reported,no comparison,very low,protective response,false,"Tesla Autopilot, according to the reported account; the vehicle model and active configuration are unverified.",,Tesla,United States,Private owner,,"Sarasota, United States",United States,Road vehicle crash avoidance,Consumer driver assistance / driver monitoring,Safety and rescue,Machines that act; Images and video,Drivers and travelers,3,High income,physical_health,,,1,1,,5,1,,personal; societal; organizational,4,2,,,,,unknown,,shown,https://en.wikipedia.org/wiki/Tesla_Autopilot,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Florida Highway Patrol troopers attribute the stop to the car's eye detection. WWSB reports that the car halted in the middle lane when the driver stopped responding to its alerts. The driver's decision to rely on Autopilot while drunk created the danger. No vehicle data were published to confirm the sequence.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 96 |
+
AIBD-0325,2025-05-24,source publication,day,unclear,A hospital's account conflicts with a relative’s claim that ChatGPT helped save a teenager from poisoning,A relative credited ChatGPT's advice to drink paraffin oil with saving a teenager from a grain pill. The medical dean at Assiut University said that the teenager showed no signs of poisoning.,Hospital disputes that ChatGPT helped save a teenager from poisoning,,,reported,Independently reported,,very low,protective response,true,"ChatGPT, according to the patient's cousin. He described the model as an advanced AI model and a limited-use trial version of GPT-5, but that identification is unverified. OpenAI publicly introduced GPT-5 on 7 August 2025, after the May report.","GPT-5 trial version, per the cousin's post; unverified",OpenAI,United States,"The patient's cousin, as a consumer user",,"Abnoub and Assiut University Hospital, Assiut Governorate, Egypt",Egypt,Healthcare: emergency toxicology,Consumer chatbot used for first-aid advice,Health,Text and language,Children; Patients,3,Lower middle income,physical_health,,,,,,,,,personal; societal,7,0,"the doctors' prognosis on arrival at the poison unit, by the cousin's account","no treatment protocol, and the case would die","discharged after 24 hours, by the cousin's account",assumed,reversed,2025-05-25,shown,https://web.archive.org/web/2026id_/https://openai.com/index/helping-people-when-they-need-it-most/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The cousin reported receiving advice on paraffin oil from ChatGPT, but the sources contain no complete transcript of the conversation and no clinical record. El Balad reproduces his post with captions of screenshots. In a direct interview with Masrawy, Dean Alaa Attia said that examinations showed no signs of phosphine poisoning and that the teenager denied swallowing the pill. The claim that AI advice caused his recovery is therefore unverified, and the hospital's account contradicts it.""}, {""label"": ""Comparison"", ""value"": ""the doctors' prognosis on arrival at the poison unit, by the cousin's account"", ""url"": ""https://www.elbalad.news/6583785""}, {""label"": ""Before"", ""value"": ""no treatment protocol, and the case would die""}, {""label"": ""After"", ""value"": ""discharged after 24 hours, by the cousin's account""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 97 |
+
AIBD-0326,2025-05-07,event,day,unclear,"Rafael's X-Guard towed decoy, labeled “AI-powered”, may have diverted Pakistani missiles from Indian Rafales in Operation Sindoor (7–10 May 2025). India's defense chief reports Indian aircraft losses. So does France's air force chief","Commentators credit Rafael's X-Guard towed decoy, labeled “AI-powered”, with diverting Pakistani missiles from Indian Rafales. No official has confirmed its use. India admitted jet losses.",X-Guard decoy may have diverted Pakistani missiles from Indian jets,,,reported,Independently reported,,very low,protective response,false,X-Guard fiber-optic towed decoy from Rafael Advanced Defense Systems. The manufacturer describes its integration with an aircraft’s electronic warfare system but does not identify an AI model or confirm use in Operation Sindoor.,,Rafael Advanced Defense Systems (Israel); several outlets misname it 'Rafale Advanced Defense Systems',Israel,Indian Air Force (claimed by commentary; unconfirmed by any official source),India,"India-Pakistan border airspace, Operation Sindoor, 7-10 May 2025",India | Pakistan,Defense: electronic warfare,Airborne towed radar decoys,Defense,,,,Lower middle income,physical_health,,,,,,,,,personal; organizational,10,0,,,,,reversed,2025-05-31,not shown,https://www.rafael.co.il/wp-content/uploads/2024/03/X-Guard-Brochure-English-03-2024.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""In its brochure, Rafael describes X-Guard as an electronic decoy integrated with the aircraft. It identifies no AI model and does not confirm deployment in Operation Sindoor. The claims of AI and of protection in this operation remain commentary, and the claim of no losses conflicts with the admission of losses by India's defense chief.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 98 |
+
AIBD-2210,2024-06-11,source publication,day,intended,Equinoct: Integrating Community Data with Flood Forecasting for a Climate-Resilient Future,"Equinoct, an Indian company, joins rainfall and river readings that residents collect with AI to forecast floods in two river basins of the Western Ghats. No forecast result is reported.",*Equinoct* builds flood forecasts from villagers' rain gauges in Kerala,,,reported,Independently reported,,very low,protective response,false,Equinoct community flood forecasting,,Equinoct,India,Equinoct with communities in two river basins,India,"Kerala, India",India,Flood forecasting with community rainfall data,,Environment,Prediction from data,People at risk of disaster,13,Lower middle income,physical_health; environmental_impact,Natural disaster prediction,,,,,,,,societal,2,0,,,,,unknown,,not shown,,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models forecast floods from rainfall and river readings. Local bodies would decide on warnings.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 99 |
AIBD-0371,2024-05-01,event,month,reported,"Within a week of tying up with Meta to route flags of self-harm from Facebook and Instagram to Abhay Command Centre, Kota police say that they prevented a possible suicide by a Jhunjhunu student","Kota police receive Meta's red-flag tags for posts about self-harm on Facebook and Instagram. Within a week, police said that they had prevented a possible suicide by a Jhunjhunu student.",Kota police act on a self-harm flag from Meta within a week of tie-up,,,reported,Independently reported,no comparison,very low,protective response,false,"Meta suicide and self-injury detection on Facebook and Instagram, with red-flag tags shared with Kota police and Abhay Command Centre",,Meta Platforms,United States,"Kota City Police (Rajasthan Police), Abhay Command Centre",India,"Kota, India",India,Suicide prevention / online safety,Social media platforms and police,Safety and rescue,Prediction from data,Students; Public servants,3,Lower middle income,emotional_health; physical_health,Emotional analysis/intelligence,,,,,5,,,personal,4,2,,,,,unknown,,shown,https://about.fb.com/news/2018/09/inside-feed-suicide-prevention-and-ai/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""PTI reports that the chain starts with a red-flag tag on Meta's system when a student posts suicide or self-harm content. Account information is then shared with police. PTI does not say whether Meta's classifier or a user report raised the flag in the Jhunjhunu case, or what happened on arrival. In its own description (2018, updated 2021), Meta says that machine-learning detection is followed by human review before authorities are contacted. The intervention itself was police work. PTI also notes that no red-flag tags had been reported in Kota city since the launch.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 100 |
AIBD-0113,2022-05-26,event,day,reported,"A threat alert from Meta, relayed through the FBI, brings Paraguayan police with psychologists to a bullied 17-year-old who praised the Uvalde shooter","Meta flagged a post in which a Paraguayan teenager praised the Uvalde shooter. After the FBI relayed the alert, police visited him with psychologists.","After Meta flags a post, police visit a Paraguayan teen",,,reported,Independently reported,no comparison,very low,protective response,false,Meta (Facebook/Instagram) violence-threat detection and law-enforcement referral process,,Meta Platforms,United States,Meta Platforms,United States,Paraguay,Paraguay,Online safety and threat referral to law enforcement,"Media, social platforms, and marketing",Safety and rescue,Prediction from data,Children,16,Upper middle income,physical_health; emotional_health,Crime reduction or prevention,,1,1,,4,1,,personal; societal,7,1,,,,,unknown,,shown,https://transparency.meta.com/enforcement/detecting-violations/how-enforcement-technology-works/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Officials said that the network detected the comment, classified it as dangerous and reported it. The report went to the FBI and to a 24/7 channel that prosecutors and the cybercrime police share. Alarcón also spoke of corroborating “datos de analítica artificial” before the visit. Neither the officials nor Meta says that the flag was automated. Meta says that its technology detects most violating content before anyone reports it, and that people review its referrals to law enforcement. Paraguayan police, prosecutors and the child-protection office identified the user, visited his home and recommended psychological support.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 101 |
AIBD-0111,2022-01-05,event,day,reported,Google's Android Earthquake Alerts reached phones in Alanya seconds before a quake of magnitude 5.1 in the Mediterranean on 5 January 2022,"Google's Android Earthquake Alerts System sent Android users near Alanya, Turkey, a warning seconds before a magnitude 5.1 offshore quake on 5 January 2022.",Android alerts reach phones in Alanya seconds before an offshore quake,Hürriyet put the lead at about 10 seconds,,reported,Independently reported,no comparison,very low,protective response,false,Android Earthquake Alerts System (crowdsourced phone-accelerometer detection; alerts for magnitude 4.5 and above),,Google,United States,Google,United States,"Alanya, Turkey",Turkey,Disaster risk reduction / earthquake early warning,Consumer smartphones (Android),Humanitarian,Prediction from data,People at risk of disaster,11,Upper middle income,environmental_impact,Natural disaster prediction,,,1,,4,1,,personal,16,2,"Feeling the shaking with no advance notice, as before the system reached Turkey",No warning before the shaking,"A phone alert about 10 seconds before the shaking, by Hürriyet's account; other outlets gave up to one minute",assumed,challenged,2023-07-27,not shown,https://blog.google/products/android/earthquake-detection-and-alerts/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Google's system produced the alert from start to finish. Its servers combined signals from the accelerometers of many phones to detect the quake and estimate its size. They then pushed alerts to phones where shaking was expected, ahead of the slower S-waves. Google does not use the words AI or machine learning when it describes this process. In a paper in Science in 2025, the authors describe it as algorithms that detect earthquakes with smartphones. Reporters give different lead times (seconds, about 10 seconds, about one minute), and Google promises seconds.""}, {""label"": ""Comparison"", ""value"": ""Feeling the shaking with no advance notice, as before the system reached Turkey"", ""url"": ""https://www.hurriyet.com.tr/gundem/erken-uyari-sistemi-tartismasi-depremin-olacagini-10-saniye-onceden-bilmek-guvenli-mi-41976168""}, {""label"": ""Before"", ""value"": ""No warning before the shaking""}, {""label"": ""After"", ""value"": ""A phone alert about 10 seconds before the shaking, by Hürriyet's account; other outlets gave up to one minute""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
|
|
|
| 134 |
AIBD-1725,2020-01-01,event,year,reported,Logically Intelligence,"Logically says that its AI service for tracking misinformation found 40,000 threats to election integrity for one US state in 2020 and raised violation detection by over 200% for one Indian state.","*Logically Intelligence* flags 40,000 election threats for a US state",200%,,claimed,The makers' claim,no comparison,very low,protective response,false,Logically Intelligence,,Logically,United Kingdom,Logically,United Kingdom,United States and India,United States | India,Government monitoring of online misinformation,,Public sector,Text and language,Public servants,16,Lower middle income; High income,social_improvement,Fact checking,2,,,,3,,,organizational,1,0,,,,,unknown,,shown,https://ircai.org/top100/entry/logically-intelligence/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Machine learning finds and groups suspect content and accounts. Analysts investigate, and clients decide whether to flag content to online services.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 135 |
AIBD-1748,2026-11-01,source access,month,reported,Real-Time Threat Detection,"Rainforest Connection says that its solar-powered Guardian devices, which use AI to find chainsaws, gunshots and engines in forest sound, have protected over 750,000 hectares of rainforest to date.","Rainforest Connection says that its devices guard 750,000 hectares",,,claimed,The makers' claim,no comparison,very low,protective response,false,Guardian (Rainforest Connection),,Rainforest Connection,United States,Rainforest Connection and local partners,,Rainforests in several countries (not listed on the page),,Forest conservation (acoustic detection of logging and poaching),,Environment,Speech and sound,,15,,environmental_impact,Environmental impacts,1,,,,3,,,societal,1,0,,,,,unknown,,shown,https://rfcx.org/guardian,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The system finds the sounds of threats and sends alerts. Local partners decide whether and how to respond.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 136 |
AIBD-0359,2026-09-01,event,month,reported,"Warning signs triggered by AI software that detects wildlife in thermal camera feeds slow drivers by an average of 3 mph on an elk corridor near Grundy, Virginia","In a Virginia pilot, Crosstek software detected wildlife in thermal camera feeds with 97 percent accuracy. Drivers slowed by an average of 3 mph at the warning signs that it activated.",Drivers slowed down when signs warned of wildlife ahead,an average speed reduction of 3 mph as vehicles approached activated signs,,claimed,The deployer's account,no method given,very low,protective response,false,"Animal Detection and Driver Warning System (ADDWS), using Crosstek thermal-camera AI detection",,Crosstek (cameras and detection software); system integration and evaluation by Virginia Tech Transportation Institute,United States,Virginia Department of Transportation with Virginia Tech Transportation Institute,United States,"Route 460 (Corridor Q) near Grundy, Buchanan County, Virginia, USA",United States,Road safety and wildlife-vehicle collision prevention,State transportation agency research pilot,Environment,Images and video,Drivers and travelers,11,High income,physical_health; animal_health,Hazard reduction,1,,4,,4,5,,personal; organizational; societal,4,2,a baseline that Virginia Tech does not specify,,an average speed reduction of 3 mph as vehicles approached signs that had been activated by wildlife detections,measured,unknown,,shown,https://news.vt.edu/articles/2026/10/vtti-elk-live-driver-warning-detection-virginia.html,"$80,771.49 per elk crash","Estimated comprehensive societal cost of an elk-vehicle crash in 2020 U.S. dollars, including economic loss and monetized pain and suffering. Context only: no AI-attributed savings or crashes prevented are established.",True,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""AI analysis of thermal camera feeds triggered the warning signs. Virginia Tech reports a 3 mph average slowdown as drivers approached activated signs, but it gives no result on collisions prevented. The results come from the evaluating institution's own report. In a separate interview with VDOT, WCYB covered the deployment and added no independent test of the measured results.""}, {""label"": ""Comparison"", ""value"": ""a baseline that Virginia Tech does not specify"", ""url"": ""https://news.vt.edu/articles/2026/10/vtti-elk-live-driver-warning-detection-virginia.html""}, {""label"": ""After"", ""value"": ""an average speed reduction of 3 mph as vehicles approached signs that had been activated by wildlife detections""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
| 137 |
+
AIBD-2190,2026-06-01,event,month,mixed,The night Venezuela shook: How Open Mapping Data Made the Emergency Response Faster,"The Humanitarian OpenStreetMap Team used its fAIr tool to rank areas and predict damage after Venezuela's earthquake of June 2026. In a test, mapping by hand was faster and more accurate.","*fAIr* ranked areas after Venezuela's quake, and a test favors hand work",,,claimed,The makers' claim,no comparison,very low,protective response,false,fAIr,,Humanitarian OpenStreetMap Team,,Humanitarian OpenStreetMap Team and its volunteer mappers,Venezuela,Venezuela (Caracas and La Guaira),Venezuela,Humanitarian mapping (building footprints and damage from satellite images),,Humanitarian,Images and video,Other workers,11,Lower middle income,operational_improvement; social_improvement,"Faster data management, decision-making",1,,,,2,,,organizational,2,0,Mapping by hand in the JOSM editor,86.08% of buildings mapped correctly by hand,78.22% of buildings mapped correctly with fAIr,measured,unknown,,shown,https://www.hotosm.org/en/news/the-night-venezuela-shook-how-open-mapping-data-made-the-emergency-response-faster/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models predict building footprints, density and damage from images. Volunteers correct and confirm every prediction before it is used.""}, {""label"": ""Comparison"", ""value"": ""Mapping by hand in the JOSM editor"", ""url"": ""https://media.ccc.de/v/sotm2025-osm-science-72051-user-testing-of-ai-assisted-mapping-tool-fair""}, {""label"": ""Before"", ""value"": ""86.08% of buildings mapped correctly by hand""}, {""label"": ""After"", ""value"": ""78.22% of buildings mapped correctly with fAIr""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
| 138 |
AIBD-0205,2026-04-01,source publication,month,reported,"Southern Water's sewer monitors flag a forming fatberg in Rowner, Gosport, so a crew clears it before any spill into the River Alver","Southern Water's machine-learning sewer monitors flagged a forming fatberg in Rowner, Gosport. A crew cleared it before any overflow into the River Alver.","Sensors flag a Gosport fatberg, which a crew clears before a spill",,,claimed,The deployer's account,no comparison,very low,protective response,false,"Southern Water's AI-enabled digital sewer system: a network of around 34,000 sewer level monitors in blockage hotspots with machine learning that learns normal sewer behavior and flags a forming blockage",,,,Southern Water,United Kingdom,"Gosport, United Kingdom",United Kingdom,Wastewater networks / pollution prevention,Regulated water and wastewater utility,Energy and infrastructure,Prediction from data,Other workers,6,High income,environmental_impact,Environmental impacts,1,1,1,,5,5,,societal; organizational,4,3,the previous reactive practice of attending blockages after a flooding or pollution event,turning up after the event to clean up and commiserate with devastated customers,spotting hundreds of potential blockages before it is too late,assumed,persisted,2026-10-08,shown,https://www.dailyecho.co.uk/news/26045031.fatberg-found-gosport-removed-causing-damage/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The level monitors detected an abnormal flow at Rowner, and the machine-learning model raised the alert to the control room. By the company's account, it distinguishes a forming blockage from daily rushes and rainfall. The jetting crew found and removed the fatberg. The articles rest on a Southern Water statement and one quoted manager. Reporters do not name the developer of the model and give no sensor data. No independent party confirmed that an overflow was imminent.""}, {""label"": ""Comparison"", ""value"": ""the previous reactive practice of attending blockages after a flooding or pollution event"", ""url"": ""https://www.dailyecho.co.uk/news/26045031.fatberg-found-gosport-removed-causing-damage/""}, {""label"": ""Before"", ""value"": ""turning up after the event to clean up and commiserate with devastated customers""}, {""label"": ""After"", ""value"": ""spotting hundreds of potential blockages before it is too late""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 139 |
AIBD-1670,2026-02-01,source publication,month,reported,AI-Driven Environmental Health Risk Prediction in Urban Morocco: The OCT7 SI Framework in Meknès,"A researcher at Moulay Ismail University says that OCT7 SI, which predicts periods of health risk in Meknès from weather, pollution and hospital admissions, supported earlier coordination in a pilot.",*OCT7 SI* warns Meknès health services of pollution and heat risk,,,claimed,The makers' claim,no comparison,very low,protective response,false,OCT7 SI (Operational Climate and Health Threat Intelligence System),,Moulay Ismail University of Meknès,Morocco,Moulay Ismail University of Meknès,Morocco,"Meknès, Morocco",Morocco,Health (surveillance of heat and air pollution risk),,Health,Prediction from data,,3; 13,Lower middle income,physical_health,Hazard reduction,1,4,,,3,,,organizational; personal,1,0,,,,,unknown,,shown,https://d19ob9sqegt2wc.cloudfront.net/stage/uploads/AI_Health_Compendium_2026_5604af298c.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models identify conditions that precede more hospital admissions. Health authorities decide on advisories and resources.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 140 |
AIBD-1673,2026-02-01,source publication,month,mixed,Spatiotemporal Probabilistic Encoder-fed Climate-aware Transfer-learning-ready Risk-forecasting Adversarial (SPECTRA) Architecture for Climate-informed Infectious Disease Forecasting,"Malaria No More's institute says that SPECTRA, a forecast model for climate-sensitive diseases, runs in national early warning systems in Indonesia and Bangladesh. No effect on outbreaks is shown.",*SPECTRA* forecasts feed disease warnings in Indonesia and Bangladesh,,,claimed,The makers' claim,no comparison,very low,protective response,false,SPECTRA,,"Institute for Health Modeling and Climate Solutions (Malaria No More), with the Indian Institute of Information Technology Vadodara",United States,Health authorities of Indonesia and Bangladesh,"Multiple: Indonesia, Bangladesh",Indonesia and Bangladesh,Indonesia | Bangladesh,Health (infectious disease early warning),,Health,Prediction from data,Public servants,3; 13,Lower middle income; Upper middle income,physical_health,Hazard reduction,1,,,,2,,,organizational; personal,1,0,,,,,unknown,,shown,https://d19ob9sqegt2wc.cloudfront.net/stage/uploads/AI_Health_Compendium_2026_5604af298c.pdf,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The model forecasts disease risk with uncertainty bounds. Public health teams decide on alerts and action.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
|
|
|
| 158 |
AIBD-1955,2021-01-01,source publication,year,reported,Wildfire Spread Prediction,"Fion, a U.S. startup, says that its deep learning maps wildfire risk, finds fire perimeters in satellite images and forecasts spread over 24 hours, for two large insurers and a U.S. state.",*Fion* forecasts wildfire spread for two insurers and a U.S. state,,,claimed,The makers' claim,no comparison,very low,protective response,false,Fion wildfire intelligence platform,,Fion Technologies Inc.,United States,Fion Technologies Inc.,United States,United States,United States,Disaster risk (wildfire spread forecasting),,Humanitarian,Images and video; Prediction from data,,13; 15,High income,knowledge_discovery; physical_health,Natural disaster prediction,1,2,,,3,,,organizational,1,0,,,,,unknown,,shown,https://ircai.org/top100/entry/wildfire-spread-prediction/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models map risk, find fire perimeters and forecast spread. Insurers and officials decide what to do.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 159 |
AIBD-1709,2019-01-01,event,year,reported,The ITIKI project combines AI with indigenous knowledge to predict droughts,"The GSMA reports that ITIKI, a drought prediction tool in South Africa, Kenya and Mozambique, sends forecasts to farmers by text message and claims 98% accuracy up to 19 months ahead.",*ITIKI* joins sensors and indigenous signs to warn farmers of drought,98%,,claimed,The makers' claim,no comparison,very low,protective response,false,ITIKI,,Central University of Technology (ITIKI project),South Africa,Central University of Technology (ITIKI project),South Africa,"South Africa, Kenya and Mozambique",South Africa | Kenya | Mozambique,Agriculture (drought forecasts for farmers),,Agriculture and food,Prediction from data,Farmers,2; 13,Low income; Lower middle income; Upper middle income,environmental_impact,Natural disaster prediction,1,,4,,3,,,personal,1,0,,,,,unknown,,not shown,,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The algorithms combine sensor readings with indigenous signs into forecasts. Farmers choose what to plant.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 160 |
AIBD-0329,2025-08-28,event,day,reported,"WINGS says, quoting the mayor, that wi.breathe cameras alert authorities to a Troizina wildfire",WINGS says that the AI fire detector on its wi.breathe cameras alerted authorities to a wildfire at Troizina on 28 August 2025. The municipality did not mention them in its statement that day.,wi.breathe cameras reportedly flag a Troizina wildfire,,,claimed,The makers' claim,no method given,very low,protective response,false,"wi.breathe Operational Intelligence Platform (WINGS ICT Solutions): hybrid optical-thermal cameras, a WINGS-trained AI smoke and fire detection model, cloud infrastructure and 5G/4G links with dashboard, email and SMS alerts",,"WINGS ICT Solutions, Athens, Greece",Greece,"Municipality of Troizinia-Methana (cameras at Vigla on Poros, Tsepeleni Aderon and Agios Georgios Methanon)",Greece,"Troizinia-Methana and Poros, Attica, Greece",Greece,Civil protection: wildfire early detection,Camera-based AI smoke and fire detection,Safety and rescue,Images and video,Public servants,13,High income,environmental_impact,Natural disaster prediction,,,1,,4,1,,societal; organizational,5,1,"the same fire without the camera alert, as the mayor describes it",,the fire was contained before it spread out of control,assumed,scaled,2026-05-12,shown,https://www.wings-ict-solutions.eu/wings-ict-solutions-delivers-wi-breathe-for-early-wildfire-detection-and-enhanced-environmental-intelligence-in-troizinia-methana-poros/,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""WINGS reports that its system detected the fire through the Methana cameras and issued real-time alerts, and the mayor credits the system in WINGS's release. In its own statement on the day, the municipality credits more than 80 firefighters, six aircraft in action within 30 minutes, the police and volunteers. It does not mention the cameras. Neither WINGS nor the municipality has given the time of the AI alert or compared it with the first call to the fire service. The AI's share in the response is therefore unknown.""}, {""label"": ""Comparison"", ""value"": ""the same fire without the camera alert, as the mayor describes it"", ""url"": ""https://www.wings-ict-solutions.eu/early-wildfire-detection-in-trizina-by-wings-wi-breathe-innovation-in-action/""}, {""label"": ""After"", ""value"": ""the fire was contained before it spread out of control""}, {""label"": ""Basis of the comparison"", ""value"": ""Assumed""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""assumed""}"
|
| 161 |
+
AIBD-2192,2025-07-09,source publication,day,intended,Meet the winners for the 2025 AI for Good Impact Awards,SidLabs built Defensa+AI without charge for the Peruvian Society for Environmental Law. It is meant to alert communities and authorities to illegal logging and mining. No result is reported.,*Defensa+AI* is built to warn Peru's Amazon defenders of illegal logging,,,claimed,The makers' claim,,very low,protective response,false,Defensa+AI,,"SidLabs, with the Sociedad Peruana de Derecho Ambiental (SPDA)",India,Sociedad Peruana de Derecho Ambiental (SPDA),Peru,Peruvian Amazon,,Environmental protection (alerts on illegal logging and mining),,Environment,,The public; Public servants,15; 16,,environmental_impact; liberty_protection,Crime reduction or prevention,,,,,,,,societal,2,0,,,,,unknown,,not shown,,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The tool is described as sending alerts on suspected illegal activity. Organizations, communities and authorities would decide how to respond.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 162 |
+
AIBD-2219,2024-12-16,source publication,day,intended,Similie: Developing Open-Source Machine Learning Models to Transform Early Warning Systems and Forecast Climate Risks,"Similie runs flood warning sensors for over 300,000 people in Timor-Leste. With UNICEF's fund it built an open machine learning model to forecast rain. The model has no reported result.",*Similie* adds a learning model to Timor-Leste's flood warnings,,,claimed,The deployer's account,,very low,protective response,false,Similie flood forecasting model (Parabl),,Similie,Timor-Leste,Similie with the Government of Timor-Leste,Timor-Leste,Timor-Leste,Timor-Leste,Flood early warning (forecast of heavy rain from local data),,Environment,Prediction from data,People at risk of disaster,13,Lower middle income,physical_health; environmental_impact,Natural disaster prediction,,,,,,,,societal,3,0,,,,,unknown,,shown,https://www.unicefventurefund.org/story/similie-developing-open-source-machine-learning-models-transform-early-warning-systems-and,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The model forecasts heavy rain from local data. The warning system and its officers send alerts.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 163 |
AIBD-2214,2024-08-01,event,month,tested,Turning Local Data into Climate Action: How Map & Rank Is Empowering Communities in Cameroon,"Map & Rank, a Cameroonian company, built a machine learning model that predicts floods and droughts for communities in the far north. The company reports an accuracy of 60%.",Map & Rank's flood model for northern Cameroon is right 60% of the time,an accuracy of 60%,,claimed,The makers' claim,no comparison,very low,protective response,false,Residat,,Map & Rank,Cameroon,Map & Rank,Cameroon,"Mayo-Danay, Far North Region, Cameroon",Cameroon,Climate risk (prediction of floods and droughts for communities),,Environment,Prediction from data,People at risk of disaster,13,Lower middle income,physical_health; environmental_impact,Natural disaster prediction,,,,,,,,societal,1,0,,,,,unknown,,shown,https://www.unicefventurefund.org/story/turning-local-data-climate-action-how-map-rank-empowering-communities-cameroon,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The model predicts floods and droughts for an area. Communities and local bodies would decide how to prepare.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 164 |
+
AIBD-2205,2024-06-17,source publication,day,intended,AEDES: A platform for dengue prediction using climate and health data for epidemic management,"Cirrolytix, a Philippine company, built AEDES to forecast dengue from climate, health and search data and to show likely hotspots from satellite images. No use by health offices is reported.",*AEDES* is built to forecast dengue hotspots in the Philippines,,,claimed,The makers' claim,,very low,protective response,false,Project AEDES,,Cirrolytix,Philippines,Cirrolytix,Philippines,Philippines,Philippines,Public health (forecasts of dengue and maps of likely hotspots),,Health,Prediction from data; Images and video,Health workers; Children,3; 13,Upper middle income,physical_health; knowledge_discovery,Hazard reduction,,,,,,,,societal,2,0,,,,,unknown,,shown,https://www.unicefventurefund.org/story/cirrolytixs-journey-revolutionize-dengue-prediction-aedes,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""The models forecast dengue counts and mark likely hotspots. Health workers would decide where to act.""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""undocumented""}"
|
| 165 |
AIBD-0036,2023-12-27,event,day,reported,"An automated classifier on a glider detects right whale calls, which prompt a voluntary NOAA Slow Zone east of Portland, Maine","The Low-Frequency Detection and Classification System on a glider detected right whales east of Portland, Maine, on 27 December 2023. NOAA declared a voluntary Slow Zone.",NOAA asks mariners to slow after glider software detects right whales,route around the area or transit at 10 knots or less,,claimed,The deployer's account,no comparison,very low,protective response,false,DMON instrument running the Low-Frequency Detection and Classification System (LFDCS) on a Slocum glider,,Woods Hole Oceanographic Institution (Mark Baumgartner),United States,Woods Hole Oceanographic Institution and NOAA Northeast Fisheries Science Center; NOAA Fisheries issues Slow Zones,United States,"Gulf of Maine east of Portland, United States",United States,Marine conservation / shipping,Endangered-species vessel-strike mitigation,Environment,Speech and sound,,14,High income,animal_health; environmental_impact,Environmental impacts,,,2,,5,1,,societal,6,2,aerial surveillance,,more slowdowns than aerial surveillance by a factor of two to five during fall and summer,measured,scaled,2025-01-21,shown,https://dcs.whoi.edu/lfdcs_manual/lfdcs_manual.html,,,,[],Adyt,"Protective response, prevention unproven.","[{""label"": ""Contribution of the AI"", ""value"": ""Automated classification produced the candidate detections. Analysts validated them on shore before NOAA acted. The resulting Slow Zone is voluntary.""}, {""label"": ""Comparison"", ""value"": ""aerial surveillance"", ""url"": ""https://tos.org/oceanography/article/glider-surveillance-for-near-real-time-detection-and-spatial-management-of-north-atlantic-right-whales""}, {""label"": ""After"", ""value"": ""more slowdowns than aerial surveillance by a factor of two to five during fall and summer""}, {""label"": ""Basis of the comparison"", ""value"": ""Measured""}]","{""category"": ""Protective response, prevention unproven"", ""comparison"": ""measured""}"
|
adyt.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
chanticleer.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
chanticleer.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:89e61355a5bf424cbc24c5acc2c34c08685c2f643ac0e7f04ca3acb51524c986
|
| 3 |
+
size 10271172
|
datapackage.json
CHANGED
|
@@ -37,8 +37,8 @@
|
|
| 37 |
"format": "csv",
|
| 38 |
"mediatype": "text/csv",
|
| 39 |
"encoding": "utf-8",
|
| 40 |
-
"bytes":
|
| 41 |
-
"hash": "sha256:
|
| 42 |
"schema": {
|
| 43 |
"fields": [
|
| 44 |
{
|
|
@@ -324,8 +324,8 @@
|
|
| 324 |
"format": "json",
|
| 325 |
"mediatype": "application/json",
|
| 326 |
"encoding": "utf-8",
|
| 327 |
-
"bytes":
|
| 328 |
-
"hash": "sha256:
|
| 329 |
},
|
| 330 |
{
|
| 331 |
"name": "chanticleer",
|
|
@@ -336,8 +336,8 @@
|
|
| 336 |
"format": "csv",
|
| 337 |
"mediatype": "text/csv",
|
| 338 |
"encoding": "utf-8",
|
| 339 |
-
"bytes":
|
| 340 |
-
"hash": "sha256:
|
| 341 |
"schema": {
|
| 342 |
"fields": [
|
| 343 |
{
|
|
@@ -623,8 +623,8 @@
|
|
| 623 |
"format": "json",
|
| 624 |
"mediatype": "application/json",
|
| 625 |
"encoding": "utf-8",
|
| 626 |
-
"bytes":
|
| 627 |
-
"hash": "sha256:
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"name": "collections-json",
|
|
@@ -704,8 +704,8 @@
|
|
| 704 |
"format": "csv",
|
| 705 |
"mediatype": "text/csv",
|
| 706 |
"encoding": "utf-8",
|
| 707 |
-
"bytes":
|
| 708 |
-
"hash": "sha256:
|
| 709 |
"schema": {
|
| 710 |
"fields": [
|
| 711 |
{
|
|
@@ -949,8 +949,8 @@
|
|
| 949 |
"format": "csv",
|
| 950 |
"mediatype": "text/csv",
|
| 951 |
"encoding": "utf-8",
|
| 952 |
-
"bytes":
|
| 953 |
-
"hash": "sha256:
|
| 954 |
"schema": {
|
| 955 |
"fields": [
|
| 956 |
{
|
|
@@ -1489,8 +1489,8 @@
|
|
| 1489 |
"format": "json",
|
| 1490 |
"mediatype": "application/json",
|
| 1491 |
"encoding": "utf-8",
|
| 1492 |
-
"bytes":
|
| 1493 |
-
"hash": "sha256:
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"name": "querist",
|
|
@@ -1501,8 +1501,8 @@
|
|
| 1501 |
"format": "csv",
|
| 1502 |
"mediatype": "text/csv",
|
| 1503 |
"encoding": "utf-8",
|
| 1504 |
-
"bytes":
|
| 1505 |
-
"hash": "sha256:
|
| 1506 |
"schema": {
|
| 1507 |
"fields": [
|
| 1508 |
{
|
|
@@ -1788,8 +1788,8 @@
|
|
| 1788 |
"format": "json",
|
| 1789 |
"mediatype": "application/json",
|
| 1790 |
"encoding": "utf-8",
|
| 1791 |
-
"bytes":
|
| 1792 |
-
"hash": "sha256:
|
| 1793 |
},
|
| 1794 |
{
|
| 1795 |
"name": "ratings",
|
|
@@ -1860,8 +1860,8 @@
|
|
| 1860 |
"format": "csv",
|
| 1861 |
"mediatype": "text/csv",
|
| 1862 |
"encoding": "utf-8",
|
| 1863 |
-
"bytes":
|
| 1864 |
-
"hash": "sha256:
|
| 1865 |
"schema": {
|
| 1866 |
"fields": [
|
| 1867 |
{
|
|
@@ -2118,8 +2118,8 @@
|
|
| 2118 |
"format": "json",
|
| 2119 |
"mediatype": "application/json",
|
| 2120 |
"encoding": "utf-8",
|
| 2121 |
-
"bytes":
|
| 2122 |
-
"hash": "sha256:
|
| 2123 |
},
|
| 2124 |
{
|
| 2125 |
"name": "requirements-json",
|
|
@@ -2166,8 +2166,8 @@
|
|
| 2166 |
"format": "csv",
|
| 2167 |
"mediatype": "text/csv",
|
| 2168 |
"encoding": "utf-8",
|
| 2169 |
-
"bytes":
|
| 2170 |
-
"hash": "sha256:
|
| 2171 |
"schema": {
|
| 2172 |
"fields": [
|
| 2173 |
{
|
|
@@ -2453,8 +2453,8 @@
|
|
| 2453 |
"format": "json",
|
| 2454 |
"mediatype": "application/json",
|
| 2455 |
"encoding": "utf-8",
|
| 2456 |
-
"bytes":
|
| 2457 |
-
"hash": "sha256:
|
| 2458 |
}
|
| 2459 |
]
|
| 2460 |
}
|
|
|
|
| 37 |
"format": "csv",
|
| 38 |
"mediatype": "text/csv",
|
| 39 |
"encoding": "utf-8",
|
| 40 |
+
"bytes": 370189,
|
| 41 |
+
"hash": "sha256:46b3a378638ed98fa822d9acff6ed8cb63a2f5514d441313e98e6bc7784f3426",
|
| 42 |
"schema": {
|
| 43 |
"fields": [
|
| 44 |
{
|
|
|
|
| 324 |
"format": "json",
|
| 325 |
"mediatype": "application/json",
|
| 326 |
"encoding": "utf-8",
|
| 327 |
+
"bytes": 9887959,
|
| 328 |
+
"hash": "sha256:3770d07489a41cedbb8eb67643bd99f1c9be462b1056c86cbb5f2c37602c4703"
|
| 329 |
},
|
| 330 |
{
|
| 331 |
"name": "chanticleer",
|
|
|
|
| 336 |
"format": "csv",
|
| 337 |
"mediatype": "text/csv",
|
| 338 |
"encoding": "utf-8",
|
| 339 |
+
"bytes": 342483,
|
| 340 |
+
"hash": "sha256:302e5597c5f2f886ce71f30e7728ac8774e2b69f1b9bcf2f6bc79816fdfd49d3",
|
| 341 |
"schema": {
|
| 342 |
"fields": [
|
| 343 |
{
|
|
|
|
| 623 |
"format": "json",
|
| 624 |
"mediatype": "application/json",
|
| 625 |
"encoding": "utf-8",
|
| 626 |
+
"bytes": 10271172,
|
| 627 |
+
"hash": "sha256:89e61355a5bf424cbc24c5acc2c34c08685c2f643ac0e7f04ca3acb51524c986"
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"name": "collections-json",
|
|
|
|
| 704 |
"format": "csv",
|
| 705 |
"mediatype": "text/csv",
|
| 706 |
"encoding": "utf-8",
|
| 707 |
+
"bytes": 330914,
|
| 708 |
+
"hash": "sha256:738d14448666db5ed2d3f84ceb87062338e1bf57ca15a05e4f081be367b4eed3",
|
| 709 |
"schema": {
|
| 710 |
"fields": [
|
| 711 |
{
|
|
|
|
| 949 |
"format": "csv",
|
| 950 |
"mediatype": "text/csv",
|
| 951 |
"encoding": "utf-8",
|
| 952 |
+
"bytes": 1188341,
|
| 953 |
+
"hash": "sha256:85e85c93a3d5ac105fd1e57c8e444eb71bb5dad43992c5ca50a7ad541a31e681",
|
| 954 |
"schema": {
|
| 955 |
"fields": [
|
| 956 |
{
|
|
|
|
| 1489 |
"format": "json",
|
| 1490 |
"mediatype": "application/json",
|
| 1491 |
"encoding": "utf-8",
|
| 1492 |
+
"bytes": 63849730,
|
| 1493 |
+
"hash": "sha256:140a113e87682473fdbc11cadbf43ef8b389bfe7658d36041bd14675cbb666ac"
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"name": "querist",
|
|
|
|
| 1501 |
"format": "csv",
|
| 1502 |
"mediatype": "text/csv",
|
| 1503 |
"encoding": "utf-8",
|
| 1504 |
+
"bytes": 1123052,
|
| 1505 |
+
"hash": "sha256:585b45110e770f78e6f37161d575f18584d90c64e954a7f36125426021327c14",
|
| 1506 |
"schema": {
|
| 1507 |
"fields": [
|
| 1508 |
{
|
|
|
|
| 1788 |
"format": "json",
|
| 1789 |
"mediatype": "application/json",
|
| 1790 |
"encoding": "utf-8",
|
| 1791 |
+
"bytes": 21817151,
|
| 1792 |
+
"hash": "sha256:bf2b452db84d557e4c8574c296d85a084ebd9e93b2ed9a593f14bd8f3dd66c68"
|
| 1793 |
},
|
| 1794 |
{
|
| 1795 |
"name": "ratings",
|
|
|
|
| 1860 |
"format": "csv",
|
| 1861 |
"mediatype": "text/csv",
|
| 1862 |
"encoding": "utf-8",
|
| 1863 |
+
"bytes": 2387733,
|
| 1864 |
+
"hash": "sha256:af795b78586f1e0860c2439c2a8c958554f0f3ea37d12b6a64edb2fecf58769f",
|
| 1865 |
"schema": {
|
| 1866 |
"fields": [
|
| 1867 |
{
|
|
|
|
| 2118 |
"format": "json",
|
| 2119 |
"mediatype": "application/json",
|
| 2120 |
"encoding": "utf-8",
|
| 2121 |
+
"bytes": 80674146,
|
| 2122 |
+
"hash": "sha256:28947ea2eada0aff91ff3dd214748abf20afc8beba658df53bd590ebe33794c6"
|
| 2123 |
},
|
| 2124 |
{
|
| 2125 |
"name": "requirements-json",
|
|
|
|
| 2166 |
"format": "csv",
|
| 2167 |
"mediatype": "text/csv",
|
| 2168 |
"encoding": "utf-8",
|
| 2169 |
+
"bytes": 577755,
|
| 2170 |
+
"hash": "sha256:2a962a7f13d0cdd0e8018a2c2fddfcc081a2f5dd80a43531f332431d06bc1345",
|
| 2171 |
"schema": {
|
| 2172 |
"fields": [
|
| 2173 |
{
|
|
|
|
| 2453 |
"format": "json",
|
| 2454 |
"mediatype": "application/json",
|
| 2455 |
"encoding": "utf-8",
|
| 2456 |
+
"bytes": 15455095,
|
| 2457 |
+
"hash": "sha256:a794d3374a028644f6852374cc52ce0217799efeaf44ac5dc65e30be9791d673"
|
| 2458 |
}
|
| 2459 |
]
|
| 2460 |
}
|
for-clinical-use.csv
CHANGED
|
@@ -24,7 +24,7 @@ AIBD-2301,Yelp review model flags San Francisco restaurants with violations,"Uni
|
|
| 24 |
AIBD-0049,Exomiser ranks gene variants to help diagnose rare disease in England,Queen Mary University of London (Exomiser team led by Damian Smedley),Exomiser phenotype-driven variant prioritization (machine-learning-assisted scoring and ranking),"Genomics England and NHS Genomic Medicine Centers; England, United Kingdom",tested against a standard,not checked,"Genetic diagnoses were found in 25% of probands, rising to 35% for monogenic disorders",Independently studied,"Expanded, 2024-12-18","The pilot's diagnostic yield and clinical actionability apply to the combined workflow of sequencing, panels, prioritization and clinical review. They cannot all be attributed to Exomiser. The authors do not say whether Exomiser or the applied panels found the variants in the two patient cases that they describe. Changes in clinical decisions do not show better health or survival across the pilot.",https://petervartanian.xyz/aibd/benefit/AIBD-0049.html
|
| 25 |
AIBD-0058,Superpower Glass briefly raises social scores of children with autism,"Stanford University (Wall lab, Division of Systems Medicine and Department of Computer Science)",Superpower Glass: Google Glass with a real-time machine-learning facial emotion classifier and companion smartphone app,"Participating families at home under the Stanford studies; Palo Alto, United States",randomized trial,not checked,gained 4.58 points more on the Vineland socialization subscale than controls (P = .005),Independently studied,"Challenged, 2019-05-02","The reports of eye contact and the change in scores in the feasibility study had no controls. The trial's advantage concerns one of four primary measures and the whole home intervention, including family interaction. Spectrum reported that the Vineland gain had disappeared six weeks after treatment ended, and the authors of the 2024 review flag low adherence and potential funding bias.",https://petervartanian.xyz/aibd/benefit/AIBD-0058.html
|
| 26 |
AIBD-2230,Kaiser's Advance Alert Monitor is tied to fewer deaths after alerts,Kaiser Permanente Northern California,Advance Alert Monitor,"Kaiser Permanente Northern California (19 hospitals); Northern California, United States",controlled comparison,not checked,an adjusted relative risk of 0.84,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2230.html
|
| 27 |
-
AIBD-0296,"Embrace wristband alerts a mother, who calls paramedics",Empatica,"Empatica Embrace wrist-worn seizure detection (electrodermal activity and accelerometer, on-board algorithm) with the Alert smartphone app","Fulton, Missouri, United States",
|
| 28 |
AIBD-0793,QVCAD cut reading time for breast ultrasound by a third,"QView Medical, Inc.",QVCAD System,QView Medical (reader study with 18 radiologists on 185 automated breast ultrasound cases); United States,controlled comparison,not checked,a 71-second or 33% reduction,Found by a regulator,"Expanded, 2017-12-21","The reader study was retrospective and used 185 existing screening exams from 2009 and 2010, of which 52 showed cancer. Readers worked in a setting built to mimic clinical reading, not in practice. Accuracy was slightly lower with QVCAD, though within the margin. Gains in patient throughput and wider use of breast ultrasound are FDA's expectations, not measurements.",https://petervartanian.xyz/aibd/benefit/AIBD-0793.html
|
| 29 |
AIBD-2231,Dascena's sepsis prediction algorithm cut hospital deaths in a trial,Dascena,Dascena machine learning severe sepsis prediction algorithm,"University of California, San Francisco Medical Center; San Francisco, California",randomized trial,not checked,from 21.3% to 8.96%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2231.html
|
| 30 |
AIBD-0931,EyeArt could halve retinal grading and miss no high-risk retinopathy,"Eyenuk, Inc.",EyeArt,"Three NHS diabetic eye screening programs in England (North East London, South East London and Gloucestershire), in a research evaluation run in parallel with usual care; North East London, South East London and Gloucestershire, England, United Kingdom",tested against a standard,not checked,"15,091 of 30,405, would still have needed human grading",Independently reported,not in the sources,"EyeArt ran in parallel with usual care, so no grader workload or patient outcome was measured. The reference standard was the human grade under the national protocol. The test covered one software version at three English programs in early 2017, and about half of the episodes would still have needed human grading.",https://petervartanian.xyz/aibd/benefit/AIBD-0931.html
|
|
@@ -104,11 +104,11 @@ AIBD-1079,CoronaryDoc met its targets for triage of stenosis in China's review,"
|
|
| 104 |
AIBD-2257,iSpy photo app improves carbohydrate counting in youth with diabetes,"Inventors at the Hospital for Sick Children, Toronto, and Western University",iSpy (carbohydrate counting app with image recognition),"Hospital for Sick Children, Toronto (trial site); Toronto, Ontario, Canada",randomized trial,not checked,not in the sources,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2257.html
|
| 105 |
AIBD-1078,MIDS-PNAS aided the detection of lung nodules on CT in China's review,Deepwise (杭州深睿博联科技有限公司),MIDS-PNAS (肺结节CT影像辅助检测软件),"Deepwise (prospective randomized crossover registration trial); Qualified medical institutions that the report does not name, for a registration in China",randomized trial,not checked,not in the sources,Found by a regulator,not in the sources,"The report gives no figures for sensitivity, specificity or reading time, and the trial was the maker's own. The reference standard was a consensus reading by three senior chest radiologists. CMDE wrote that long-term safety and effectiveness still need tracking and asked for further study of how the algorithm generalizes.",https://petervartanian.xyz/aibd/benefit/AIBD-1078.html
|
| 106 |
AIBD-0735,Zio XT patch catches more heart rhythm problems than a 24-hour Holter,iRhythm Technologies,Zio XT (Zio biosensor with ZEUS deep neural network ECG analysis),"NHS organizations in England, with studies that included UK NHS sites; England, United Kingdom (guidance), with studies in the United Kingdom and United States",controlled comparison,not checked,"In three of four comparative studies, the yield over total wear time was higher",Found by an assessment body,"Expanded, 2023-07-18","The four comparative studies had 357 people in all, too varied for a meta-analysis. In the UK trial, 20% refused the Holter monitor, and the assessors judged it likely underpowered. No study clearly defined accuracy against standard care, and the company's accuracy data were per episode, not per patient. No evidence shows that the higher yield improves clinical outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0735.html
|
| 107 |
-
AIBD-0001,A brain–spine interface restored control of walking after paralysis,NeuroRestore (EPFL/CHUV/UNIL) with CEA-Clinatec and ONWARD Medical,STIMO-BSI brain–spine interface,"Lausanne University Hospital (CHUV); Lausanne, Switzerland",
|
| 108 |
AIBD-0638,BoneView helps Poitiers emergency doctors read fracture X-rays,Gleamer,BoneView,"Poitiers University Hospital (CHU de Poitiers); Emergency department, Poitiers University Hospital, Poitiers, France",before and after,"3,720 adult trauma patients with limb radiographs (1,780 in 2020, 1,940 in 2021) at one level I trauma center.",the rate of discrepancies between emergency physicians' and radiologists' readings fell by 17%,Independently studied,not in the sources,"The study is a retrospective before-and-after comparison at one hospital. The authors note that other, uncontrolled changes could explain the shorter stay. The change in clinically relevant discrepancies was not significant. Consultation counts came from medical reports that likely missed informal calls. The authors did not assess the software's financial impact.",https://petervartanian.xyz/aibd/benefit/AIBD-0638.html
|
| 109 |
AIBD-0775,NaviCam ProScan helped doctors find more lesions in less time,"Ankon Technologies Co., Ltd.",NaviCam ProScan,"Seven European centers in the prospective ARTIC study (22 capsule endoscopy readers); Seven centers in Italy, France, Germany, Hungary, Spain, Sweden and the United Kingdom",controlled comparison,not checked,3 minutes and 50 seconds per video,Found by a regulator,"Persisted, 2026-10-11","The study enrolled 133 patients at seven European centers and did not collect race or ethnicity. Performance of the device without clinician input was not evaluated, so standalone performance was not established, and the number of false positive predictions from the software alone is unknown. FDA requires postmarket testing to show how the device performs in use.",https://petervartanian.xyz/aibd/benefit/AIBD-0775.html
|
| 110 |
AIBD-1069,EA PC-LGI helped endoscopists find more polyps in a Chinese trial,"Wuhan EndoAngel Medical Technology Co., Ltd. (武汉楚精灵医疗科技有限公司)",EA PC-LGI (肠息肉电子下消化道内窥镜图像辅助检测软件),"Wuhan EndoAngel Medical Technology (prospective trial at Ruijin Hospital, Beijing Cancer Hospital and the Central Hospital of Wuhan); Shanghai, Beijing and Wuhan, China",controlled comparison,not checked,"detected 84.97% of polyps, against 72.07% in colonoscopies without it",Found by a regulator,not in the sources,"EndoAngel sponsored the trial and supplied EA PC-LGI. Endoscopists knew when the software was on, and the authors cannot rule out bias. The overall polyp detection rate did not differ between the groups. The primary outcomes were endoscopist sensitivity and specificity, so the trial did not measure cancer outcomes. The trial ran at three large centers in China.",https://petervartanian.xyz/aibd/benefit/AIBD-1069.html
|
| 111 |
-
AIBD-1690,Meza cuts reporting time by 20% for health workers in Turkana clinics,World Food Programme and Charitable Analytics International,Meza,"World Food Programme; Turkana County, Kenya",
|
| 112 |
AIBD-1723,"Rwanda's Covid-19 chatbot reaches 150,000 users in its first 10 days","GIZ FAIR Forward, with local partners and the Mozilla Foundation",Covid-19 chatbot for Rwanda (FAIR Forward),"GIZ FAIR Forward, with the Government of Rwanda; Rwanda",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1723.html
|
| 113 |
AIBD-1726,MedCheX reads chest X-rays for pneumonia for staff in 61 countries,National Cheng Kung University,MedCheX,National Cheng Kung University; Taiwan (users in 61 countries),no comparison,not checked,0.893,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1726.html
|
| 114 |
AIBD-1759,ECHealth is built to find drug effects in the ECGs of trial patients,SoftMining,ECHealth,"SoftMining; Salerno, Italy",not checked,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1759.html
|
|
@@ -124,7 +124,7 @@ AIBD-1936,Guarding Beats app calls for help on a gesture or a fall,EvoAId Ltd.,E
|
|
| 124 |
AIBD-1943,ESA and UNICEF's AI forecasts dengue in Peru one month ahead in a study,"European Space Agency Φ-Lab, with UNICEF",ESA Φ-Lab dengue forecasting ensemble,"European Space Agency Φ-Lab, with UNICEF; Peru",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1943.html
|
| 125 |
AIBD-1944,RobHome voice assistant supports elderly people in 20 Spanish homes,Yasyt Robotics SL and Andorra Research + Innovation Foundation,RobHome,"Yasyt Robotics SL and Andorra Research + Innovation Foundation; Barcelona and Olot, Spain",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1944.html
|
| 126 |
AIBD-1952,Arogya World plans AI to pick influencers for healthy eating in India,Arogya World,MyThali influence optimization,Arogya World; India,not checked,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1952.html
|
| 127 |
-
AIBD-2169,Crisis Text Line's model brings high-risk texters to counselors sooner,Crisis Text Line,Risk triage model for incoming texts (unnamed),Crisis Text Line; United States,
|
| 128 |
AIBD-2209,"Elsa supports diagnoses for 5,000 Tanzanian patients a month",Inspired Ideas (Elsa Health),Elsa Health Assistant,Drug dispensaries and health facilities in Tanzania; Tanzania,no comparison,not checked,"roughly 5,000 patients a month",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2209.html
|
| 129 |
AIBD-2236,AI stroke decision support cuts new vascular events in a Chinese trial,Beijing Tiantan Hospital and the China National Clinical Research Center for Neurological Diseases,Stroke clinical decision support system (stroke CDSS) of the GOLDEN BRIDGE II trial,38 hospitals in China (the trial's intervention group); China,randomized trial,not checked,from 3.9% to 2.9%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2236.html
|
| 130 |
AIBD-2311,Epic Deterioration Index alerts precede fewer escalations in care,"Epic Systems Corporation (the index), with an alert workflow set up at the hospital",Epic Deterioration Index alerts with a shared workflow of nurses and physicians,One academic hospital in the United States (four general internal medicine units); United States,controlled comparison,not checked,10.4 percentage points,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2311.html
|
|
@@ -233,17 +233,17 @@ AIBD-2306,Opioid Halo cut needless oxygen alarms for people who take opioids,Mas
|
|
| 233 |
AIBD-0080,Five early cancers are found after qXR flags nodules on unrelated X-rays,Qure.ai,qXR,"Participating centers of the CREATE study, not named in the paper; the clinical authors are affiliated with Hacettepe University (Ankara) and with Aarthi Scans and Sri Ramachandra Medical College and Hospital (Chennai)",no comparison,not checked,"confirmed early-stage lung cancer in five adults, and four had surgery",The deployer's account,not in the sources,"The authors of the series do not say whether radiologists found the same nodules on their own. So earlier detection by AI, comparative rates of false positives and any survival benefit remain unresolved.",https://petervartanian.xyz/aibd/benefit/AIBD-0080.html
|
| 234 |
AIBD-2250,EchoNet heart-function estimates need fewer corrections than humans',Researchers at Stanford University and Cedars-Sinai Medical Center,EchoNet-Dynamic (video-based deep learning model for left ventricular ejection fraction),"Echocardiography laboratory of Cedars-Sinai Medical Center; Los Angeles, California, United States",randomized trial,not checked,16.8%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2250.html
|
| 235 |
AIBD-0887,CT scans came six days sooner after Annalise Enterprise CXR arrived,Annalise.ai,Annalise Enterprise CXR,"NHS Grampian; Grampian, Scotland, United Kingdom",before and after,not checked,received a CT scan six days more quickly,Found by an assessment body,"Persisted, 2025-08-21","SHTG compared 68 patients after the change with a 2019 baseline of 113, and says it cannot attribute the differences to the AI because other changes came at the same time. It found limited or no published evidence on effectiveness, cost effectiveness or safety. The evaluation did not capture long-term patient outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0887.html
|
| 236 |
-
AIBD-1068,MY-CJ-01 raised detection of polyps on a first pass in Tencent's trial,"Tencent Healthcare (Shenzhen) Co., Ltd. (腾讯医疗健康(深圳)有限公司)",MY-CJ-01 (结肠息肉电子内窥镜图像辅助检测软件),"Tencent Healthcare (Shenzhen) (prospective randomized trial with 296 patients at two centers); Two trial centers that the report does not name, for a registration in China",
|
| 237 |
AIBD-0776,KidneyIntelX.dkd sorted diabetic patients by risk to their kidneys,"Renalytix AI, Inc.",KidneyIntelX.dkd,Renalytix (retrospective validation on biobanked samples from 657 adults); A single biorepository drawing on 10 clinical sites (not named in the decision summary),no comparison,not checked,"5.9%, 21.5% and 66.9%",Found by a regulator,"Persisted, 2026-04-22","Renalytix, the maker, ran the validation as a retrospective analysis of stored samples from one biorepository. Follow-up was incomplete, so the five-year rates are Kaplan-Meier estimates. Only 55 adults fell in the high level. The submission held no information on patient perspectives. Anthem reports that the evidence comes largely from retrospective and observational studies and shows no meaningful gain in patient-centered outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0776.html
|
| 238 |
AIBD-1140,EU guideline group backs AI-supported double reading of mammograms,"Three vendors that asked to remain anonymous in the study behind the estimate, with ScreenPoint Medical (Transpara), Lunit (INSIGHT MMG), Kheiron Medical Technologies (Mia) and DeepHealth in the other studies reviewed","AI support for double reading of screening mammograms: three commercial algorithms kept anonymous in the study behind the estimate, with Transpara, Lunit INSIGHT MMG, Mia and a DeepHealth model in the other studies reviewed","Screening services in the studies that the ECIBC guideline group reviewed, with Karolinska University Hospital's archived examinations behind the estimate and organized screening programs in the EU as intended users; European Union (guideline), with the study behind the estimate in Stockholm, Sweden",tested against a standard,not checked,"one more breast cancer for every 1,000 women screened",Found by an assessment body,"Persisted, 2026-11-06",The ECIBC group rated the evidence as of very low certainty because the accuracy of the readings is uncertain and no data show effects on breast cancer or mortality. No study used the readings in a screening program or followed women over time. The group calls for prospective studies that measure outcomes such as breast cancer and mortality.,https://petervartanian.xyz/aibd/benefit/AIBD-1140.html
|
| 239 |
AIBD-2060,"India's Just Ask! chatbot reaches 140,000 users on sexual health",United Nations Population Fund (UNFPA),Just Ask! (Khulke Poocho!) chatbot,"UNFPA India, with the state governments of Madhya Pradesh and Rajasthan; India (launched in Madhya Pradesh, with outreach in 22 states)",no comparison,not checked,"140,000",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2060.html
|
| 240 |
-
AIBD-0736,Lung screening centers may use AI nodule detection under NICE guidance,"Siemens Healthineers, Coreline Soft, Riverain Technologies, contextflow, Infervision, Arterys, Qure.ai, SenseTime, MeVis, Aidence and VUNO","AI-derived computer-aided detection software for lung nodules: AI-Rad Companion Chest CT, AVIEW LCS+, ClearRead CT, contextflow SEARCH Lung CT, InferRead CT Lung, Lung AI, qCT-Lung, SenseCare-Lung Pro, Veolity, Veye Lung Nodules and VUNO Med-LungCT AI","Screening centers in the studies reviewed by NICE and NHS targeted lung cancer screening centers in England during evidence generation; England, United Kingdom (guidance), with studies in several countries",
|
| 241 |
AIBD-1067,EASYASPECT brought doctors' scores of stroke CT closer to experts',"Yukun (Beijing) Network Technology Co., Ltd. (语坤(北京)网络科技有限公司)",EASYASPECT (缺血性卒中CT图像辅助评估软件),"Yukun (Beijing) Network Technology (retrospective reader trial at the Second Affiliated Hospital of Nanchang University and Hwa Mei Hospital of the University of Chinese Academy of Sciences in Ningbo); Ningbo, Zhejiang and Nanchang, Jiangxi, China",controlled comparison,not checked,"a kappa of 0.80 with EASYASPECT, against 0.55 without it",Found by a regulator,not in the sources,"The trial was retrospective, with 224 past scans, two hospitals and 8 doctors. Its main outcome was agreement with an expert panel, not patient outcomes. CMDE says that the software cannot serve alone as a basis for treatment decisions and must not be used in emergency care.",https://petervartanian.xyz/aibd/benefit/AIBD-1067.html
|
| 242 |
AIBD-0631,Viz HCM picked out a hard-to-diagnose heart disease from ECGs,"Viz.ai, Inc.",Viz HCM,"Viz.ai (retrospective study on ECGs from three Massachusetts hospitals); Boston area and Salem, Massachusetts, United States",no comparison,"3,196 ECGs (291 with and 2,905 without hypertrophic cardiomyopathy) from three Massachusetts hospitals, 2017 to 2022.",a sensitivity of 68.4% and a specificity of 99.1%,Found by a regulator,"Persisted, 2026-03-23","The study used historical patient data from three Massachusetts hospitals and involved no patients. A single cardiologist labeled each ECG, and a second cardiologist rechecked only 60. Usability testing involved cardiologists, not non-specialist clinicians. The submission did not include patient perspectives.",https://petervartanian.xyz/aibd/benefit/AIBD-0631.html
|
| 243 |
AIBD-0119,A surgical robot helps a surgeon remove a brain clot from an awake man,Synaptive Medical Inc. (Toronto),"Modus V Synaptive robotic digital microscope with Neuro-Navigation AI (image fusion, fiber-tract display, surgical simulation, trajectory warnings)","Tam Anh General Hospital, Ho Chi Minh City; Ho Chi Minh City, Vietnam",no comparison,not checked,About 40 ml of clot came out in just over 30 minutes,The deployer's account,"Persisted, 2024-12-23","Several teams took part in the operation, so its outcome cannot be credited to the software alone. The hospital gives no comparison of this patient's outcome with conventional surgery, and no long-term neurological follow-up is reported. In its account, the hospital calls the gap between a 14:00 arrival and a 17:00 start two hours.",https://petervartanian.xyz/aibd/benefit/AIBD-0119.html
|
| 244 |
AIBD-1090,DrBrain-AD flagged hippocampal changes on brain MRI in China's review,"Shenzhen Yiwei Medical Technology Co., Ltd. (深圳市铱硙医疗科技有限公司)",DrBrain-AD (磁共振图像辅助评估软件),"Shenzhen Yiwei Medical Technology (registration trial with 514 patients at three hospitals); Xuanwu Hospital of Capital Medical University, the Second Xiangya Hospital of Central South University and the First Affiliated Hospital of Wenzhou Medical University, China",tested against a standard,not checked,a sensitivity of 0.9328 and a specificity of 0.9042,Found by a regulator,not in the sources,"The trial was single-arm and multicenter, run by the applicant, with 514 patients at three hospitals. There was no unaided reading arm, so the report gives no figures for doctors reading without the software. The targets came from pre-clinical data and guideline consensus.",https://petervartanian.xyz/aibd/benefit/AIBD-1090.html
|
| 245 |
AIBD-0803,AI100 with Shonit matched manual microscopy on 882 blood smears,SigTuple Technologies Pvt. Ltd.,AI100 with Shonit,SigTuple (method comparison against manual microscopy at four sites),tested against a standard,not checked,92.7% sensitivity and 95.4% specificity for abnormal white cells,Found by a regulator,not in the sources,"The study measured agreement with manual microscopy at four sites, not accuracy against a clinical diagnosis or any effect on care. It did not measure time saved. Agreement was weakest for monocytes, with a correlation of 0.789. Results depend on a skilled operator, who identifies and classifies each cell.",https://petervartanian.xyz/aibd/benefit/AIBD-0803.html
|
| 246 |
-
AIBD-0654,NICE finds AI outlining quicker for radiotherapy planning,"Siemens Healthineers, TheraPanacea, Mirada Medical, Carina Medical, Limbus AI, MIM Software, Philips, MVision AI and RaySearch","AI autocontouring software: AI-Rad Companion Organs RT, ART-Plan, DLCExpert, INTContour, Limbus Contour, MIM Contour ProtégéAI, MRCAT Prostate plus Auto-contouring, MVision Segmentation Service and RayStation","Radiotherapy centers in the studies reviewed by NICE; NHS radiotherapy services in England; Multiple countries (studies); England, United Kingdom (guidance)",
|
| 247 |
AIBD-1669,AI tools in 28 Dhaka pharmacies cut symptom-based antibiotic sales,"United International University's AIMS Lab, with CMED Health Limited",AI-supported antibiotic stewardship system for community pharmacies,"United International University's AIMS Lab, with CMED Health Limited; Dhaka, Bangladesh",before and after,not checked,6.42%,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1669.html
|
| 248 |
AIBD-0641,A phone camera coach improved blood pressure in hypertension,RaphAI Health Technologies Limited,Rehabilitation Exercise Prescription app (AI pose recognition built on TensorFlow and OpenPose),"Zhujiang Hospital, Southern Medical University, and Guangdong Provincial People's Hospital; Guangzhou, Guangdong, China",randomized trial,"62 adults aged 18 to 75 with hypertension, Guangzhou.",The app group gained 62.77 meters more in the 6-minute walk test than the control group,Independently studied,not in the sources,"The trial enrolled 62 patients by convenience sampling at two Guangzhou hospitals, was open-label and lasted eight weeks. The authors say that the small sample from one region limits generalizability and that eight weeks cannot show long-term effects.",https://petervartanian.xyz/aibd/benefit/AIBD-0641.html
|
| 249 |
AIBD-1066,EC01-01 helped endoscopists find more colon polyps in a trial,"Xiamen Yingnuo Medical Technology Co., Ltd. (厦门影诺医疗科技有限公司)",EC01-01 (肠息肉电子下消化道图像辅助检测软件),"Xiamen Yingnuo Medical Technology (prospective parallel-controlled trial with 1,020 subjects at four sites); Four trial sites in China that the report does not name",controlled comparison,not checked,"detected 94.7% of polyps, against 84.5% without it",Found by a regulator,"Expanded, 2025-05-22","The trial was the maker's own, a prospective parallel-controlled study at four Chinese sites, and the report does not say whether patients were randomized. The reference came from two experts reading the recorded videos, with a third settling disagreements. CMDE says the output is for reference only and no clinical decision may rest on it alone.",https://petervartanian.xyz/aibd/benefit/AIBD-1066.html
|
|
@@ -278,7 +278,7 @@ AIBD-0582,A nurse returns to full-time work after software detects her brain ble
|
|
| 278 |
AIBD-0770,Samsung's Sleep Apnea Feature caught most moderate to severe cases,"Samsung Electronics Co., Ltd.",Sleep Apnea Feature (Samsung Galaxy Watch and Phone),Samsung (multi-center clinical validation against polysomnography run by an accredited sleep lab); Sites not named in the decision summary,tested against a standard,not checked,sensitivity was 82.7%,Found by a regulator,"Expanded, 2025-06-05","The study enrolled an enriched population, and 47 of 620 enrolled subjects did not finish. The feature gave no usable result on 16.7% of nights. The specificity target was met only after a post-hoc analysis. Subgroups were not powered for conclusive comparisons. FDA says that 10 people were directed to care but reports no follow-up.",https://petervartanian.xyz/aibd/benefit/AIBD-0770.html
|
| 279 |
AIBD-2263,Machine learning sepsis alert speeds antibiotics in a Bangkok trial,Not stated in the abstract. The authors work at Chulalongkorn University.,Sepsis alert by a machine learning model on a staff display (King Chulalongkorn Memorial Hospital),"Emergency department of King Chulalongkorn Memorial Hospital, Bangkok; Bangkok, Thailand",randomized trial,not checked,68.4%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2263.html
|
| 280 |
AIBD-1065,Dr.Wise-AADS helped doctors find more brain aneurysms on CT,Hangzhou Deepwise (杭州深睿博联科技有限公司),Dr.Wise-AADS (颅内动脉瘤CT造影图像辅助检测软件),"Hangzhou Deepwise (retrospective crossover reader trial on 965 cases at four institutions); Four trial institutions that the report does not name, for a registration in China",controlled comparison,not checked,lesion-level sensitivity for brain aneurysms on CT angiograms 10.85% higher,Found by a regulator,not in the sources,"The trial was the applicant's own retrospective crossover study. Doctors read stored scans, and imaging experts' readings were the reference. It measured sensitivity and specificity, not patient outcomes. The software assists detection of only some aneurysms, those of 3 mm or more.",https://petervartanian.xyz/aibd/benefit/AIBD-1065.html
|
| 281 |
-
AIBD-1161,EndoBRAIN-EYE wins a Japanese insurance add-on for better colonoscopy,"Cybernet Systems Co., Ltd. (サイバネットシステム株式会社), with Showa University Northern Yokohama Hospital and Nagoya University",EndoBRAIN-EYE (内視鏡画像診断支援プログラム),Japanese hospitals and clinics where colonoscopy specialists use it during polyp removal (not named); Japan,
|
| 282 |
AIBD-0742,Wales supports CADe colonoscopy after it found more adenomas in trials,"Medtronic (GI Genius), Olympus (ENDO-AID), Fujifilm (CAD EYE), Wision AI (EndoScreener), Wuhan EndoAngel Medical Technology Company (EndoAngel) and other makers","Computer-aided detection (CADe) colonoscopy systems in the trials that HTW pooled, including GI Genius, ENDO-AID, CAD EYE, EndoScreener and EndoAngel","Endoscopy services in the pooled randomized trials; Multiple countries (trials), with guidance for Wales, United Kingdom",randomized trial,not checked,23% higher with CADe than with standard colonoscopy,Found by an assessment body,not in the sources,"The main source of bias was that endoscopists and outcome assessors were not blinded. HTW reports high heterogeneity and wide confidence intervals. It found no evidence that CADe raised detection of advanced adenomas or of carcinoma, which was rare. No study reported subsequent surgery, time to diagnosis or quality of life. HTW asked for data on real-world use.",https://petervartanian.xyz/aibd/benefit/AIBD-0742.html
|
| 283 |
AIBD-2284,DAX Copilot trims note time for University of Pennsylvania clinicians,Nuance Communications,DAX Copilot,"University of Pennsylvania Health System (outpatient practices); Philadelphia, Pennsylvania, United States",before and after,not checked,from 10.3 to 8.2 minutes,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2284.html
|
| 284 |
AIBD-0626,Sepsis ImmunoScore sorted patients by sepsis risk in FDA's review,"Prenosis, Inc.",Sepsis ImmunoScore,"Prenosis (retrospective validation on data from three US hospitals); Boston, Massachusetts; Chicago, Illinois; Royal Oak, Michigan, United States",tested against a standard,746 patients with suspected sepsis at three US hospitals.,an area under the ROC curve of 0.81,Found by a regulator,"Expanded, 2026-07-28","The study was retrospective, using data collected prospectively at three hospitals, so the score guided no care. Physicians at the patients' own hospitals judged who had sepsis, though FDA recommends independent adjudicators at separate institutions. Only 33 patients fell in the very high group, and its 95% interval for sepsis ran from 51.29% to 84.41%.",https://petervartanian.xyz/aibd/benefit/AIBD-0626.html
|
|
@@ -394,7 +394,7 @@ AIBD-0401,A university is developing a moral-tale game for autistic children,Chr
|
|
| 394 |
AIBD-0402,A university plans software that corrects rehabilitation exercises,Anna University,AI-driven exercise assessment and live-feedback system,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"A planned evaluation cannot show that the feedback is accurate or that it improves recovery beyond usual rehabilitation. Safety, adherence and performance in unsupervised homes remain untested.",https://petervartanian.xyz/aibd/benefit/AIBD-0402.html
|
| 395 |
AIBD-0404,A Nigerian team plans software to make impaired speech clearer,AI4Africa,AI4Africa AI-powered AAC tool,Nigeria,not checked,not checked,not in the sources,The makers' claim,not in the sources,The team names Limi Hospital as one of the hospitals that it plans to approach to test the tool. The hospital was not a confirmed testbed. The team has not reported that any person with a speech disability used the tool. Gains in speech intelligibility and communication accuracy require completed development and an evaluation with participants.,https://petervartanian.xyz/aibd/benefit/AIBD-0404.html
|
| 396 |
AIBD-0405,A Tanzanian university plans wearable alerts for abnormal vital signs,Mbeya University of Science and Technology,AI-Based Health Monitoring and Advisory System,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"The team names no pilot, and fewer complications of disease are described as an expected impact. The team gives no detection performance, alert acted on, clinical outcome or comparison of access. Planned biosensors and deeper models further limit what can be said about the present system.",https://petervartanian.xyz/aibd/benefit/AIBD-0405.html
|
| 397 |
-
AIBD-0406,EquiDermAI is meant to reduce gaps between skin tones in AI diagnosis,
|
| 398 |
AIBD-0407,A journaling tool in development would give students emotional feedback,Vellore Institute of Technology,StudentSphere AI journaling system,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"In its summary, the team says that an internal pilot is planned, while in its description it calls the product actively deployed. Clinical validation and oversight by psychiatrists remain future work. The figures for accuracy in classifying emotions and for response time are targets, and no completed crisis response or measured fall in stress has been reported.",https://petervartanian.xyz/aibd/benefit/AIBD-0407.html
|
| 399 |
AIBD-0408,Researchers plan hospital tests of a wheelchair steered by brain waves,Vellore Institute of Technology,Brain-wave driven wheelchair for paralysis patients using EEG and ML,not in the sources,no comparison,not checked,not in the sources,The makers' claim,not in the sources,"Training data from ten people do not establish independent control by people with paralysis. The researchers give no participant characteristics, held-out testing, command errors, safety failures, control comparators or sustained mobility outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0408.html
|
| 400 |
AIBD-0409,A planned smartwatch system would monitor facial paralysis,"Vellore Institute of Technology, Chennai",Smart Watch based Integration System for Habituating Face Paralysis Monitoring,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"The team labels deployment active while it places mimicry-recognition models, their training data and a suitable watch in future work. No patient result resolves that discrepancy, and earlier intervention and better quality of life are prospective claims.",https://petervartanian.xyz/aibd/benefit/AIBD-0409.html
|
|
@@ -413,14 +413,14 @@ AIBD-0142,ChatGPT ties a mother's 18-month cough to her blood pressure drug,Open
|
|
| 413 |
AIBD-0656,NICE lets NHS talking therapies use the Limbic Access chatbot,Limbic,Limbic Access (AI self-referral chatbot),"NHS Talking Therapies for anxiety and depression services in England; England, United Kingdom",controlled comparison,"129,400 patients in NHS Talking Therapies services in England.","referrals rise by 15%, against 6% in control services",Found by an assessment body,Persisted,The 15% rise comes from an observational study. NICE says that most evidence came from the companies or from authors affiliated with them. It found a lack of evidence on people who could not access the service. It says that it is unclear whether the evidence applies to all NHS providers. It says that evidence on time saved is limited.,https://petervartanian.xyz/aibd/benefit/AIBD-0656.html
|
| 414 |
AIBD-0740,NICE allows the Wysa referral chatbot in NHS talking therapies,Wysa,Wysa Digital Referral Assistant (Wysa DRA),"NHS Talking Therapies for anxiety and depression services in England; England, United Kingdom",no comparison,not checked,16.0 to 21.0 minutes per assessment,Independently reported,not in the sources,"NICE's committee had only company-reported, non-comparative data for the Wysa chatbot, none from published research. It could not pool the evidence because of its heterogeneity. It found that evidence on time saved was limited and that the net time saved remains uncertain. It found the evidence's generalisability across NHS Talking Therapies services unclear and the impact on clinical decision making unclear.",https://petervartanian.xyz/aibd/benefit/AIBD-0740.html
|
| 415 |
AIBD-0633,ArteraAI Prostate sorted prostate cancer patients by risk of spread,"Artera, Inc.",ArteraAI Prostate,"Artera (clinical validation at three US sites); Three sites, United States",tested against a standard,886 patients at three US sites.,10-year metastasis risk was 28.1% in its high risk category,Found by a regulator,not in the sources,"The FDA review reports how well the categories separated 10-year risk in 886 patients. It does not show that treatment guided by the software changed outcomes, and FDA says only that the device may help reduce under- or over-treatment. Only 72 patients were African American. The submission held no information on patient perspectives.",https://petervartanian.xyz/aibd/benefit/AIBD-0633.html
|
| 416 |
-
AIBD-0064,"After a scan aided by Brainomix finds a stroke, she leaves in 25 minutes",Brainomix,Brainomix 360 Stroke,"NHS England / Royal Berkshire Hospital; Reading, United Kingdom",
|
| 417 |
AIBD-0305,ChatGPT lists a rare condition that a genetic test confirms,OpenAI,ChatGPT,"Cardiff, United Kingdom",no comparison,not checked,not in the sources,Independently reported,"Persisted, 2026-10-05","The evidence is the patient's account through two outlets. The reporters do not say that they inspected clinical records, the prompt, the model version or the full list of candidates. Diagnosis alone does not establish a treatment benefit, and reports of how long the coma lasted disagree.",https://petervartanian.xyz/aibd/benefit/AIBD-0305.html
|
| 418 |
AIBD-0134,"ChatGPT flags an insulinoma that doctors missed, which surgery removes",OpenAI,ChatGPT,"Majadahonda, Spain",no comparison,not checked,surgeons removed about two thirds of his pancreas and his symptoms went,Independently reported,"Persisted, 2025-08-07","Every report traces to the account of the patient or his partner, and none gives pathology results or testimony from a clinician who treated him. The only physician quoted, David Céspedes, had no part in his care. Some headlines spoke of pancreatic cancer, but an insulinoma is not by itself evidence of a malignant cancer.",https://petervartanian.xyz/aibd/benefit/AIBD-0134.html
|
| 419 |
AIBD-0136,Staff agree with ChatGPT that a worsening spider bite is an emergency,OpenAI,ChatGPT,"Wolfforth, United States",no comparison,not checked,not in the sources,Independently reported,not in the sources,"No confirmed diagnosis, treatment or follow-up is reported, and the clinicians' comments come only through the woman's account. This case does not show whether ChatGPT saved her life or whether its advice on where to seek care is accurate in general.",https://petervartanian.xyz/aibd/benefit/AIBD-0136.html
|
| 420 |
AIBD-0271,GPT-5 lists oncocytoma first in two retrospective cases,OpenAI,"ChatGPT (GPT-5, 2025 version)","Department of Medicine, Surgery and Pharmacy, University of Sassari; Sassari, Italy",no comparison,"Two patients with caruncular oncocytoma confirmed by histology at the University of Sassari eye clinic, plus three published images of other caruncular tumors.",not in the sources,The deployer's account,not in the sources,"The original two diagnoses were known before testing. Additional tumor tests expose misclassification, but the small selected sample does not establish sensitivity or specificity. Queries were not repeated, and the study did not compare independent clinician assessments or measure prospective patient benefit.",https://petervartanian.xyz/aibd/benefit/AIBD-0271.html
|
| 421 |
AIBD-2178,CHARM gives South African health workers care advice for each client,"Causal Foundry, with mothers2mothers",CHARM,"mothers2mothers; Western Cape, South Africa",no comparison,not checked,not in the sources,The deployer's account,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2178.html
|
| 422 |
AIBD-2176,"Ask-A-Question answers 880,000 questions from South African mothers",IDinsight and Reach Digital Health,Ask-A-Question (AAQ),"National Department of Health of South Africa (MomConnect), with Reach Digital Health; South Africa",no comparison,not checked,"more than 880,000 questions",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2176.html
|
| 423 |
-
AIBD-1091,MIAS-3000 flagged macular degeneration on OCT scans in China's review,"Suzhou Bigvision Medical Technology Co., Ltd. (苏州比格威医疗科技有限公司)",MIAS-3000 (眼科OCT图像辅助诊断软件),Suzhou Bigvision Medical Technology (prospective multicenter registration trial with 908 subjects); Trial sites in China that the report does not name,
|
| 424 |
AIBD-0739,AI tools find spinal fractures missed in routine scan reports,"Aidoc Medical (BriefCase-Triage), Avicenna.AI (CINA-VCF Quantix), Nanox AI (HealthVCF and HealthOST) and IB Lab (IB Lab FLAMINGO)","AI tools for opportunistic detection of vertebral fragility fractures on CT: BriefCase-Triage, CINA-VCF Quantix, HealthVCF, HealthOST and IB Lab FLAMINGO","Hospitals in the studies reviewed by NICE and NHS radiology services in England during the evidence generation period; England, United Kingdom (guidance), with studies in several countries",tested against a standard,not checked,a specificity above 90%,Found by an assessment body,Persisted,"Most studies of the AI tools were retrospective and measured diagnostic accuracy, and most comparisons with standard care were done outside the UK. NICE says that it is uncertain how much the tools improve detection in the NHS. Very little evidence shows how they affect clinical management. The model's standard-care values came from a small expert elicitation study.",https://petervartanian.xyz/aibd/benefit/AIBD-0739.html
|
| 425 |
AIBD-0959,AI-Rad Companion Prostate MR lifted readers' accuracy slightly,"Siemens Healthineers, with 上海西门子医疗器械有限公司 (Siemens Shanghai Medical Equipment Ltd.) as the applicant in China",AI-Rad Companion Prostate MR (前列腺癌磁共振图像辅助检测软件),"上海西门子医疗器械有限公司 (Siemens Shanghai Medical Equipment Ltd.), the applicant (retrospective reader trial with 12 doctors and 340 subjects); Trial sites that the report does not name, for a registration in China",controlled comparison,not checked,"an AUROC of 0.7010, against 0.6758 without it",Found by a regulator,"Expanded, 2025-11-07","The trial was retrospective, with 12 doctors reading images from 340 subjects. Siemens Shanghai, the applicant, supplied the trial data. The AUROC gain of 0.0252 had a 95% confidence interval from 0.0011 to 0.0493. The sensitivity gain of 4.4% had an interval from -0.2% to 9.1%. Median reading time was 98 seconds in both groups.",https://petervartanian.xyz/aibd/benefit/AIBD-0959.html
|
| 426 |
AIBD-2185,AMTI answers refugee women's health questions on WhatsApp in Türkiye,"MEDAK, with HERA Digital Health",AMTI (Predictive Care Navigator),"MEDAK, with HERA Digital Health; Southeast Türkiye",no comparison,not checked,"nearly 2,500 users",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2185.html
|
|
@@ -434,7 +434,7 @@ AIBD-0378,CA-GPT matches expert stent choices more often than junior physicians,
|
|
| 434 |
AIBD-0912,Norfolk's AI tool for preventing falls is credited with fewer admissions,Not named in the assessment,AI-driven proactive falls prevention tool,"Norfolk County Council, adult social care; Norfolk, England, United Kingdom",no comparison,not checked,not in the sources,Independently reported,"Expanded, 2025-12-12",The evidence is council leaders' account of a pilot and partner feedback. CQC says that data supported it but gives no figures and does not say whose data it saw.,https://petervartanian.xyz/aibd/benefit/AIBD-0912.html
|
| 435 |
AIBD-1033,AI-based image analysis detects referable diabetic retinopathy,"Several vendors, among them Eyenuk (EyeArt), Digital Diagnostics (IDx-DR) and EyRIS (SELENA)","AI-based image analysis of fundus photographs for diabetic retinopathy screening (products in the included studies such as EyeArt, IDx-DR, SELENA, RetCAD and Aireen)","Clinics and screening programs in the comparative studies that HTA syd reviewed; Studies in Europe, Asia, Africa, Oceania and North America, assessed for Sweden's Southern Healthcare Region",tested against a standard,not checked,about 92% sensitivity and 91% specificity,Found by an assessment body,"Persisted, 2026-04-22","HTA syd's pooled accuracy comes from 13 non-randomized studies that compared AI grading with human graders, all from other health systems. HTA syd could not judge whether recalls, future vision loss or blindness change with AI grading. Sweden's grading scale differs from the international standard, which makes the results hard to transfer. Moderate certainty means that more research is needed, and no current cost-effectiveness analysis exists.",https://petervartanian.xyz/aibd/benefit/AIBD-1033.html
|
| 436 |
AIBD-0023,An Apple Watch called for help after a runner’s heart stopped,Apple,Apple Watch Fall Detection,"Varsseveld, Gelderland, Netherlands",no comparison,not checked,began cardiopulmonary resuscitation about 6–7 minutes after the collapse,Independently studied,"Persisted, 2026-05-01","The authors give no date or place. The identification of the patient, the date and the street are inferred from interviews and a pager message. Men's Health says that the runner also pressed the SOS button himself and that a pulse returned within about three minutes. This account does not fit the timeline in the paper. The case cannot show how quickly he would have been found without the alert.",https://petervartanian.xyz/aibd/benefit/AIBD-0023.html
|
| 437 |
-
AIBD-0891,Clinicians feel less burdened by documentation with AI scribes,"Several makers, including Nuance (Dragon Ambient eXperience), Abridge, Nabla, Sunoh.ai and Amazon","AI scribes (ambient scribes), including Dragon Ambient eXperience, Abridge, Nabla Copilot, Sunoh.ai and Amazon HealthScribe","Hospitals, clinics and other services in the studies that AIHTA reviewed; Multiple countries (studies), mostly the United States",
|
| 438 |
AIBD-2276,AI scribe adopters at UCSF Health bill more work units per week,Not named in the report,Ambient AI scribes (products not named in the report),"UCSF Health (ambulatory practices); San Francisco, California, United States",controlled comparison,not checked,1.81 relative value units per week,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2276.html
|
| 439 |
AIBD-0156,Ward staff revive a patient after thynC flags a pre-arrest heart rhythm,"Seers Technology Co., Seongnam (maker of the ThynC patch); Daewoong Pharmaceutical supplies thynC","thynC (씽크) AI-based smart bed monitoring system: wearable sensors measuring ECG, heart rate and respiratory rate in real time, with alarms sent to a remote central monitor","Gwangju Hanguk Hospital (광주한국병원); Gwangju, South Korea",no comparison,not checked,heart rate climbed to 192 beats per minute,The makers' claim,"Expanded, 2026-06-19","The account comes through the supplier, with no clinical record and no response time that anyone independent has assessed. Staff resuscitated the patient, and the time that would have passed without the alert is unknown. A restored circulation leaves long-term recovery unknown.",https://petervartanian.xyz/aibd/benefit/AIBD-0156.html
|
| 440 |
AIBD-0267,Mother's AI query prompts a rare-disease diagnosis in a 22-month-old,Google (Med-Gemini-3D),Smartphone application described by the authors as powered by Med-Gemini-3D,"The patient's mother (consumer use); United States (authors at Garnet Health Medical Center, Middletown, New York)",no comparison,not checked,not in the sources,Independently studied,not in the sources,"In their abstract, the authors do not identify the app or show that it ran Med-Gemini-3D. They do not say which therapeutic interventions followed or what they achieved.",https://petervartanian.xyz/aibd/benefit/AIBD-0267.html
|
|
@@ -475,7 +475,7 @@ AIBD-2243,AMIE improves cardiologists' assessments of past complex patients,Goog
|
|
| 475 |
AIBD-0274,Lunit INSIGHT CXR flags a subpulmonic effusion that ultrasound confirms,"Lunit Inc., Seoul",Lunit INSIGHT CXR,"Basic Emergency Service, Vila Real de Santo Antonio Health Center, Algarve Local Health Unit; Vila Real de Santo Antonio, Algarve, Portugal",no comparison,not checked,not in the sources,The deployer's account,not in the sources,"Correct identification does not establish subsequent clinical improvement. The patient's refusal limits the observed outcome, and the retained case does not isolate what the radiographer would have recognized without the AI output.",https://petervartanian.xyz/aibd/benefit/AIBD-0274.html
|
| 476 |
AIBD-2277,Ambient AI scribe shortens note time in an emergency department,Nuance Communications,"Ambient AI scribe (DAX Copilot, as the outside review names it)","A tertiary academic medical center in California (the authors work at Stanford); Stanford, California, United States",controlled comparison,not checked,2 minutes 45 seconds,Independently studied,"Persisted, 2026-07-02",not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2277.html
|
| 477 |
AIBD-0957,uAI-BreastTomoCare helped doctors find more breast masses,"United Imaging Intelligence (Beijing) Co., Ltd. (联影智能医疗科技(北京)有限公司)",uAI-BreastTomoCare (乳腺X射线断层图像辅助检测软件),"United Imaging Intelligence (Beijing) (retrospective reader trial on images of 432 subjects from three centers); Three trial centers that the report does not name, for a registration in China",controlled comparison,not checked,"62.17% of breast mass lesions on tomosynthesis, against 50.42% without it",Found by a regulator,not in the sources,"The trial was the maker's own retrospective reading study of stored images from three centers. It measured mass detection, not outcomes for patients. Used alone, the software had a specificity of 53.81% at the patient level. The review says that trained doctors must use it and that it cannot serve alone as the basis for clinical decisions.",https://petervartanian.xyz/aibd/benefit/AIBD-0957.html
|
| 478 |
-
AIBD-0788,Claire OCT System helped surgeons remove cancer left at the margins,"Perimeter Medical Imaging AI, Inc.",Claire OCT System with the ImgAssist AI detection algorithm,"Breast surgeons at the investigational sites of the randomized trial (NCT05113927); Investigational sites, United States",
|
| 479 |
AIBD-0767,Tyto Insights detects bulging eardrums better than non-ENT doctors,Tyto Care Ltd.,Tyto Insights for Eardrum Bulging Detection,Tyto Care (retrospective validation on 857 otoscope recordings from real-world use); Recording sites not stated in the decision summary,tested against a standard,not checked,a sensitivity of 0.9231,Found by a regulator,"Persisted, 2026-04-27",not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0767.html
|
| 480 |
AIBD-0653,England's health technology assessor supports AI help in colonoscopy,"Odin Vision (CADDIE), Fujifilm (CAD EYE), Olympus Medical Systems (ENDO-AID), Wision AI (EndoScreener), Medtronic (GI Genius) and Magentiq Eye (MAGENTIQ-COLO)","AI computer-aided detection (CADe) for colonoscopy: CADDIE, CAD EYE, ENDO-AID, EndoScreener, GI Genius and MAGENTIQ-COLO","Endoscopy services in the trials reviewed by NICE; NHS services in England during the evidence generation period; Multiple countries (trials); England, United Kingdom (guidance)",randomized trial,"70 studies of AI-assisted live colonoscopy, mostly randomized controlled trials.",not in the sources,Found by an assessment body,not in the sources,"NICE's committee judged it uncertain whether the higher adenoma detection rate will lower colorectal cancer rates. Gains for advanced adenomas and sessile serrated lesions were not statistically significant for any system. Most studies excluded people with Lynch syndrome, familial adenomatous polyposis or inflammatory bowel disease. Data on false positives were very limited. The quality and quantity of evidence varied across systems.",https://petervartanian.xyz/aibd/benefit/AIBD-0653.html
|
| 481 |
AIBD-0741,ArtiQ.Spiro raised sensitivity for COPD in a UK trial,Clario,ArtiQ.Spiro,SPIRO-AID trial with 133 UK primary care professionals reviewing past spirometry records; United Kingdom,randomized trial,not checked,not in the sources,Found by an assessment body,not in the sources,"In SPIRO-AID, professionals reviewed 50 past spirometry tests, a small sample with few normal results. The external group calculated accuracy from data that the authors supplied in confidence. Accuracy in primary care use is unclear, and no study reported long-term outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0741.html
|
|
|
|
| 24 |
AIBD-0049,Exomiser ranks gene variants to help diagnose rare disease in England,Queen Mary University of London (Exomiser team led by Damian Smedley),Exomiser phenotype-driven variant prioritization (machine-learning-assisted scoring and ranking),"Genomics England and NHS Genomic Medicine Centers; England, United Kingdom",tested against a standard,not checked,"Genetic diagnoses were found in 25% of probands, rising to 35% for monogenic disorders",Independently studied,"Expanded, 2024-12-18","The pilot's diagnostic yield and clinical actionability apply to the combined workflow of sequencing, panels, prioritization and clinical review. They cannot all be attributed to Exomiser. The authors do not say whether Exomiser or the applied panels found the variants in the two patient cases that they describe. Changes in clinical decisions do not show better health or survival across the pilot.",https://petervartanian.xyz/aibd/benefit/AIBD-0049.html
|
| 25 |
AIBD-0058,Superpower Glass briefly raises social scores of children with autism,"Stanford University (Wall lab, Division of Systems Medicine and Department of Computer Science)",Superpower Glass: Google Glass with a real-time machine-learning facial emotion classifier and companion smartphone app,"Participating families at home under the Stanford studies; Palo Alto, United States",randomized trial,not checked,gained 4.58 points more on the Vineland socialization subscale than controls (P = .005),Independently studied,"Challenged, 2019-05-02","The reports of eye contact and the change in scores in the feasibility study had no controls. The trial's advantage concerns one of four primary measures and the whole home intervention, including family interaction. Spectrum reported that the Vineland gain had disappeared six weeks after treatment ended, and the authors of the 2024 review flag low adherence and potential funding bias.",https://petervartanian.xyz/aibd/benefit/AIBD-0058.html
|
| 26 |
AIBD-2230,Kaiser's Advance Alert Monitor is tied to fewer deaths after alerts,Kaiser Permanente Northern California,Advance Alert Monitor,"Kaiser Permanente Northern California (19 hospitals); Northern California, United States",controlled comparison,not checked,an adjusted relative risk of 0.84,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2230.html
|
| 27 |
+
AIBD-0296,"Embrace wristband alerts a mother, who calls paramedics",Empatica,"Empatica Embrace wrist-worn seizure detection (electrodermal activity and accelerometer, on-board algorithm) with the Alert smartphone app","Fulton, Missouri, United States",no comparison,not checked,not in the sources,Independently reported,"Persisted, 2026-01-07","The rescue rests on a single news account, and no clinical record of the event has been found. The same article covers a death despite a successful alert, which shows that a timely human response is a separate requirement.",https://petervartanian.xyz/aibd/benefit/AIBD-0296.html
|
| 28 |
AIBD-0793,QVCAD cut reading time for breast ultrasound by a third,"QView Medical, Inc.",QVCAD System,QView Medical (reader study with 18 radiologists on 185 automated breast ultrasound cases); United States,controlled comparison,not checked,a 71-second or 33% reduction,Found by a regulator,"Expanded, 2017-12-21","The reader study was retrospective and used 185 existing screening exams from 2009 and 2010, of which 52 showed cancer. Readers worked in a setting built to mimic clinical reading, not in practice. Accuracy was slightly lower with QVCAD, though within the margin. Gains in patient throughput and wider use of breast ultrasound are FDA's expectations, not measurements.",https://petervartanian.xyz/aibd/benefit/AIBD-0793.html
|
| 29 |
AIBD-2231,Dascena's sepsis prediction algorithm cut hospital deaths in a trial,Dascena,Dascena machine learning severe sepsis prediction algorithm,"University of California, San Francisco Medical Center; San Francisco, California",randomized trial,not checked,from 21.3% to 8.96%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2231.html
|
| 30 |
AIBD-0931,EyeArt could halve retinal grading and miss no high-risk retinopathy,"Eyenuk, Inc.",EyeArt,"Three NHS diabetic eye screening programs in England (North East London, South East London and Gloucestershire), in a research evaluation run in parallel with usual care; North East London, South East London and Gloucestershire, England, United Kingdom",tested against a standard,not checked,"15,091 of 30,405, would still have needed human grading",Independently reported,not in the sources,"EyeArt ran in parallel with usual care, so no grader workload or patient outcome was measured. The reference standard was the human grade under the national protocol. The test covered one software version at three English programs in early 2017, and about half of the episodes would still have needed human grading.",https://petervartanian.xyz/aibd/benefit/AIBD-0931.html
|
|
|
|
| 104 |
AIBD-2257,iSpy photo app improves carbohydrate counting in youth with diabetes,"Inventors at the Hospital for Sick Children, Toronto, and Western University",iSpy (carbohydrate counting app with image recognition),"Hospital for Sick Children, Toronto (trial site); Toronto, Ontario, Canada",randomized trial,not checked,not in the sources,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2257.html
|
| 105 |
AIBD-1078,MIDS-PNAS aided the detection of lung nodules on CT in China's review,Deepwise (杭州深睿博联科技有限公司),MIDS-PNAS (肺结节CT影像辅助检测软件),"Deepwise (prospective randomized crossover registration trial); Qualified medical institutions that the report does not name, for a registration in China",randomized trial,not checked,not in the sources,Found by a regulator,not in the sources,"The report gives no figures for sensitivity, specificity or reading time, and the trial was the maker's own. The reference standard was a consensus reading by three senior chest radiologists. CMDE wrote that long-term safety and effectiveness still need tracking and asked for further study of how the algorithm generalizes.",https://petervartanian.xyz/aibd/benefit/AIBD-1078.html
|
| 106 |
AIBD-0735,Zio XT patch catches more heart rhythm problems than a 24-hour Holter,iRhythm Technologies,Zio XT (Zio biosensor with ZEUS deep neural network ECG analysis),"NHS organizations in England, with studies that included UK NHS sites; England, United Kingdom (guidance), with studies in the United Kingdom and United States",controlled comparison,not checked,"In three of four comparative studies, the yield over total wear time was higher",Found by an assessment body,"Expanded, 2023-07-18","The four comparative studies had 357 people in all, too varied for a meta-analysis. In the UK trial, 20% refused the Holter monitor, and the assessors judged it likely underpowered. No study clearly defined accuracy against standard care, and the company's accuracy data were per episode, not per patient. No evidence shows that the higher yield improves clinical outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0735.html
|
| 107 |
+
AIBD-0001,A brain–spine interface restored control of walking after paralysis,NeuroRestore (EPFL/CHUV/UNIL) with CEA-Clinatec and ONWARD Medical,STIMO-BSI brain–spine interface,"Lausanne University Hospital (CHUV); Lausanne, Switzerland",before and after,"One 38-year-old man at CHUV in Lausanne, ten years after an incomplete C5/C6 spinal cord injury.",walk more than 10 meters with crutches while the stimulation was off,Independently studied,"Expanded, 2026-01-22",The team reported on one man in Nature. ONWARD has counted later implants but has not said how many recipients recovered movement. No independent study is among the sources.,https://petervartanian.xyz/aibd/benefit/AIBD-0001.html
|
| 108 |
AIBD-0638,BoneView helps Poitiers emergency doctors read fracture X-rays,Gleamer,BoneView,"Poitiers University Hospital (CHU de Poitiers); Emergency department, Poitiers University Hospital, Poitiers, France",before and after,"3,720 adult trauma patients with limb radiographs (1,780 in 2020, 1,940 in 2021) at one level I trauma center.",the rate of discrepancies between emergency physicians' and radiologists' readings fell by 17%,Independently studied,not in the sources,"The study is a retrospective before-and-after comparison at one hospital. The authors note that other, uncontrolled changes could explain the shorter stay. The change in clinically relevant discrepancies was not significant. Consultation counts came from medical reports that likely missed informal calls. The authors did not assess the software's financial impact.",https://petervartanian.xyz/aibd/benefit/AIBD-0638.html
|
| 109 |
AIBD-0775,NaviCam ProScan helped doctors find more lesions in less time,"Ankon Technologies Co., Ltd.",NaviCam ProScan,"Seven European centers in the prospective ARTIC study (22 capsule endoscopy readers); Seven centers in Italy, France, Germany, Hungary, Spain, Sweden and the United Kingdom",controlled comparison,not checked,3 minutes and 50 seconds per video,Found by a regulator,"Persisted, 2026-10-11","The study enrolled 133 patients at seven European centers and did not collect race or ethnicity. Performance of the device without clinician input was not evaluated, so standalone performance was not established, and the number of false positive predictions from the software alone is unknown. FDA requires postmarket testing to show how the device performs in use.",https://petervartanian.xyz/aibd/benefit/AIBD-0775.html
|
| 110 |
AIBD-1069,EA PC-LGI helped endoscopists find more polyps in a Chinese trial,"Wuhan EndoAngel Medical Technology Co., Ltd. (武汉楚精灵医疗科技有限公司)",EA PC-LGI (肠息肉电子下消化道内窥镜图像辅助检测软件),"Wuhan EndoAngel Medical Technology (prospective trial at Ruijin Hospital, Beijing Cancer Hospital and the Central Hospital of Wuhan); Shanghai, Beijing and Wuhan, China",controlled comparison,not checked,"detected 84.97% of polyps, against 72.07% in colonoscopies without it",Found by a regulator,not in the sources,"EndoAngel sponsored the trial and supplied EA PC-LGI. Endoscopists knew when the software was on, and the authors cannot rule out bias. The overall polyp detection rate did not differ between the groups. The primary outcomes were endoscopist sensitivity and specificity, so the trial did not measure cancer outcomes. The trial ran at three large centers in China.",https://petervartanian.xyz/aibd/benefit/AIBD-1069.html
|
| 111 |
+
AIBD-1690,Meza cuts reporting time by 20% for health workers in Turkana clinics,World Food Programme and Charitable Analytics International,Meza,"World Food Programme; Turkana County, Kenya",before and after,not checked,20%,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1690.html
|
| 112 |
AIBD-1723,"Rwanda's Covid-19 chatbot reaches 150,000 users in its first 10 days","GIZ FAIR Forward, with local partners and the Mozilla Foundation",Covid-19 chatbot for Rwanda (FAIR Forward),"GIZ FAIR Forward, with the Government of Rwanda; Rwanda",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1723.html
|
| 113 |
AIBD-1726,MedCheX reads chest X-rays for pneumonia for staff in 61 countries,National Cheng Kung University,MedCheX,National Cheng Kung University; Taiwan (users in 61 countries),no comparison,not checked,0.893,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1726.html
|
| 114 |
AIBD-1759,ECHealth is built to find drug effects in the ECGs of trial patients,SoftMining,ECHealth,"SoftMining; Salerno, Italy",not checked,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1759.html
|
|
|
|
| 124 |
AIBD-1943,ESA and UNICEF's AI forecasts dengue in Peru one month ahead in a study,"European Space Agency Φ-Lab, with UNICEF",ESA Φ-Lab dengue forecasting ensemble,"European Space Agency Φ-Lab, with UNICEF; Peru",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1943.html
|
| 125 |
AIBD-1944,RobHome voice assistant supports elderly people in 20 Spanish homes,Yasyt Robotics SL and Andorra Research + Innovation Foundation,RobHome,"Yasyt Robotics SL and Andorra Research + Innovation Foundation; Barcelona and Olot, Spain",no comparison,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1944.html
|
| 126 |
AIBD-1952,Arogya World plans AI to pick influencers for healthy eating in India,Arogya World,MyThali influence optimization,Arogya World; India,not checked,not checked,not in the sources,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1952.html
|
| 127 |
+
AIBD-2169,Crisis Text Line's model brings high-risk texters to counselors sooner,Crisis Text Line,Risk triage model for incoming texts (unnamed),Crisis Text Line; United States,no comparison,not checked,the median wait for high-risk texters was 3 minutes,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2169.html
|
| 128 |
AIBD-2209,"Elsa supports diagnoses for 5,000 Tanzanian patients a month",Inspired Ideas (Elsa Health),Elsa Health Assistant,Drug dispensaries and health facilities in Tanzania; Tanzania,no comparison,not checked,"roughly 5,000 patients a month",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2209.html
|
| 129 |
AIBD-2236,AI stroke decision support cuts new vascular events in a Chinese trial,Beijing Tiantan Hospital and the China National Clinical Research Center for Neurological Diseases,Stroke clinical decision support system (stroke CDSS) of the GOLDEN BRIDGE II trial,38 hospitals in China (the trial's intervention group); China,randomized trial,not checked,from 3.9% to 2.9%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2236.html
|
| 130 |
AIBD-2311,Epic Deterioration Index alerts precede fewer escalations in care,"Epic Systems Corporation (the index), with an alert workflow set up at the hospital",Epic Deterioration Index alerts with a shared workflow of nurses and physicians,One academic hospital in the United States (four general internal medicine units); United States,controlled comparison,not checked,10.4 percentage points,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2311.html
|
|
|
|
| 233 |
AIBD-0080,Five early cancers are found after qXR flags nodules on unrelated X-rays,Qure.ai,qXR,"Participating centers of the CREATE study, not named in the paper; the clinical authors are affiliated with Hacettepe University (Ankara) and with Aarthi Scans and Sri Ramachandra Medical College and Hospital (Chennai)",no comparison,not checked,"confirmed early-stage lung cancer in five adults, and four had surgery",The deployer's account,not in the sources,"The authors of the series do not say whether radiologists found the same nodules on their own. So earlier detection by AI, comparative rates of false positives and any survival benefit remain unresolved.",https://petervartanian.xyz/aibd/benefit/AIBD-0080.html
|
| 234 |
AIBD-2250,EchoNet heart-function estimates need fewer corrections than humans',Researchers at Stanford University and Cedars-Sinai Medical Center,EchoNet-Dynamic (video-based deep learning model for left ventricular ejection fraction),"Echocardiography laboratory of Cedars-Sinai Medical Center; Los Angeles, California, United States",randomized trial,not checked,16.8%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2250.html
|
| 235 |
AIBD-0887,CT scans came six days sooner after Annalise Enterprise CXR arrived,Annalise.ai,Annalise Enterprise CXR,"NHS Grampian; Grampian, Scotland, United Kingdom",before and after,not checked,received a CT scan six days more quickly,Found by an assessment body,"Persisted, 2025-08-21","SHTG compared 68 patients after the change with a 2019 baseline of 113, and says it cannot attribute the differences to the AI because other changes came at the same time. It found limited or no published evidence on effectiveness, cost effectiveness or safety. The evaluation did not capture long-term patient outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0887.html
|
| 236 |
+
AIBD-1068,MY-CJ-01 raised detection of polyps on a first pass in Tencent's trial,"Tencent Healthcare (Shenzhen) Co., Ltd. (腾讯医疗健康(深圳)有限公司)",MY-CJ-01 (结肠息肉电子内窥镜图像辅助检测软件),"Tencent Healthcare (Shenzhen) (prospective randomized trial with 296 patients at two centers); Two trial centers that the report does not name, for a registration in China",controlled comparison,not checked,"The polyp miss rate was 23.39%, against 44.67% in group A",Found by a regulator,not in the sources,"The trial was Tencent's own, at two centers, with 265 of 296 patients analyzed. CMDE's summary does not say how group A and group B used MY-CJ-01. The endpoint was polyp detection on a first pass, not any outcome of care. The software's marks are for reference only, and the physician must judge whether a polyp is present.",https://petervartanian.xyz/aibd/benefit/AIBD-1068.html
|
| 237 |
AIBD-0776,KidneyIntelX.dkd sorted diabetic patients by risk to their kidneys,"Renalytix AI, Inc.",KidneyIntelX.dkd,Renalytix (retrospective validation on biobanked samples from 657 adults); A single biorepository drawing on 10 clinical sites (not named in the decision summary),no comparison,not checked,"5.9%, 21.5% and 66.9%",Found by a regulator,"Persisted, 2026-04-22","Renalytix, the maker, ran the validation as a retrospective analysis of stored samples from one biorepository. Follow-up was incomplete, so the five-year rates are Kaplan-Meier estimates. Only 55 adults fell in the high level. The submission held no information on patient perspectives. Anthem reports that the evidence comes largely from retrospective and observational studies and shows no meaningful gain in patient-centered outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0776.html
|
| 238 |
AIBD-1140,EU guideline group backs AI-supported double reading of mammograms,"Three vendors that asked to remain anonymous in the study behind the estimate, with ScreenPoint Medical (Transpara), Lunit (INSIGHT MMG), Kheiron Medical Technologies (Mia) and DeepHealth in the other studies reviewed","AI support for double reading of screening mammograms: three commercial algorithms kept anonymous in the study behind the estimate, with Transpara, Lunit INSIGHT MMG, Mia and a DeepHealth model in the other studies reviewed","Screening services in the studies that the ECIBC guideline group reviewed, with Karolinska University Hospital's archived examinations behind the estimate and organized screening programs in the EU as intended users; European Union (guideline), with the study behind the estimate in Stockholm, Sweden",tested against a standard,not checked,"one more breast cancer for every 1,000 women screened",Found by an assessment body,"Persisted, 2026-11-06",The ECIBC group rated the evidence as of very low certainty because the accuracy of the readings is uncertain and no data show effects on breast cancer or mortality. No study used the readings in a screening program or followed women over time. The group calls for prospective studies that measure outcomes such as breast cancer and mortality.,https://petervartanian.xyz/aibd/benefit/AIBD-1140.html
|
| 239 |
AIBD-2060,"India's Just Ask! chatbot reaches 140,000 users on sexual health",United Nations Population Fund (UNFPA),Just Ask! (Khulke Poocho!) chatbot,"UNFPA India, with the state governments of Madhya Pradesh and Rajasthan; India (launched in Madhya Pradesh, with outreach in 22 states)",no comparison,not checked,"140,000",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2060.html
|
| 240 |
+
AIBD-0736,Lung screening centers may use AI nodule detection under NICE guidance,"Siemens Healthineers, Coreline Soft, Riverain Technologies, contextflow, Infervision, Arterys, Qure.ai, SenseTime, MeVis, Aidence and VUNO","AI-derived computer-aided detection software for lung nodules: AI-Rad Companion Chest CT, AVIEW LCS+, ClearRead CT, contextflow SEARCH Lung CT, InferRead CT Lung, Lung AI, qCT-Lung, SenseCare-Lung Pro, Veolity, Veye Lung Nodules and VUNO Med-LungCT AI","Screening centers in the studies reviewed by NICE and NHS targeted lung cancer screening centers in England during evidence generation; England, United Kingdom (guidance), with studies in several countries",controlled comparison,not checked,"In all nine studies of reading time, review was faster with the software",Found by an assessment body,Persisted,"NICE found five studies, on four products, with per-person screening results, and none compared products. Its cost model was exploratory and indicative only. The reading-time studies were partly in laboratory-like conditions. Five of the 13 products had no published evidence. A 2026 IQWiG report on AI reading of chest images found no studies of patient-relevant outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0736.html
|
| 241 |
AIBD-1067,EASYASPECT brought doctors' scores of stroke CT closer to experts',"Yukun (Beijing) Network Technology Co., Ltd. (语坤(北京)网络科技有限公司)",EASYASPECT (缺血性卒中CT图像辅助评估软件),"Yukun (Beijing) Network Technology (retrospective reader trial at the Second Affiliated Hospital of Nanchang University and Hwa Mei Hospital of the University of Chinese Academy of Sciences in Ningbo); Ningbo, Zhejiang and Nanchang, Jiangxi, China",controlled comparison,not checked,"a kappa of 0.80 with EASYASPECT, against 0.55 without it",Found by a regulator,not in the sources,"The trial was retrospective, with 224 past scans, two hospitals and 8 doctors. Its main outcome was agreement with an expert panel, not patient outcomes. CMDE says that the software cannot serve alone as a basis for treatment decisions and must not be used in emergency care.",https://petervartanian.xyz/aibd/benefit/AIBD-1067.html
|
| 242 |
AIBD-0631,Viz HCM picked out a hard-to-diagnose heart disease from ECGs,"Viz.ai, Inc.",Viz HCM,"Viz.ai (retrospective study on ECGs from three Massachusetts hospitals); Boston area and Salem, Massachusetts, United States",no comparison,"3,196 ECGs (291 with and 2,905 without hypertrophic cardiomyopathy) from three Massachusetts hospitals, 2017 to 2022.",a sensitivity of 68.4% and a specificity of 99.1%,Found by a regulator,"Persisted, 2026-03-23","The study used historical patient data from three Massachusetts hospitals and involved no patients. A single cardiologist labeled each ECG, and a second cardiologist rechecked only 60. Usability testing involved cardiologists, not non-specialist clinicians. The submission did not include patient perspectives.",https://petervartanian.xyz/aibd/benefit/AIBD-0631.html
|
| 243 |
AIBD-0119,A surgical robot helps a surgeon remove a brain clot from an awake man,Synaptive Medical Inc. (Toronto),"Modus V Synaptive robotic digital microscope with Neuro-Navigation AI (image fusion, fiber-tract display, surgical simulation, trajectory warnings)","Tam Anh General Hospital, Ho Chi Minh City; Ho Chi Minh City, Vietnam",no comparison,not checked,About 40 ml of clot came out in just over 30 minutes,The deployer's account,"Persisted, 2024-12-23","Several teams took part in the operation, so its outcome cannot be credited to the software alone. The hospital gives no comparison of this patient's outcome with conventional surgery, and no long-term neurological follow-up is reported. In its account, the hospital calls the gap between a 14:00 arrival and a 17:00 start two hours.",https://petervartanian.xyz/aibd/benefit/AIBD-0119.html
|
| 244 |
AIBD-1090,DrBrain-AD flagged hippocampal changes on brain MRI in China's review,"Shenzhen Yiwei Medical Technology Co., Ltd. (深圳市铱硙医疗科技有限公司)",DrBrain-AD (磁共振图像辅助评估软件),"Shenzhen Yiwei Medical Technology (registration trial with 514 patients at three hospitals); Xuanwu Hospital of Capital Medical University, the Second Xiangya Hospital of Central South University and the First Affiliated Hospital of Wenzhou Medical University, China",tested against a standard,not checked,a sensitivity of 0.9328 and a specificity of 0.9042,Found by a regulator,not in the sources,"The trial was single-arm and multicenter, run by the applicant, with 514 patients at three hospitals. There was no unaided reading arm, so the report gives no figures for doctors reading without the software. The targets came from pre-clinical data and guideline consensus.",https://petervartanian.xyz/aibd/benefit/AIBD-1090.html
|
| 245 |
AIBD-0803,AI100 with Shonit matched manual microscopy on 882 blood smears,SigTuple Technologies Pvt. Ltd.,AI100 with Shonit,SigTuple (method comparison against manual microscopy at four sites),tested against a standard,not checked,92.7% sensitivity and 95.4% specificity for abnormal white cells,Found by a regulator,not in the sources,"The study measured agreement with manual microscopy at four sites, not accuracy against a clinical diagnosis or any effect on care. It did not measure time saved. Agreement was weakest for monocytes, with a correlation of 0.789. Results depend on a skilled operator, who identifies and classifies each cell.",https://petervartanian.xyz/aibd/benefit/AIBD-0803.html
|
| 246 |
+
AIBD-0654,NICE finds AI outlining quicker for radiotherapy planning,"Siemens Healthineers, TheraPanacea, Mirada Medical, Carina Medical, Limbus AI, MIM Software, Philips, MVision AI and RaySearch","AI autocontouring software: AI-Rad Companion Organs RT, ART-Plan, DLCExpert, INTContour, Limbus Contour, MIM Contour ProtégéAI, MRCAT Prostate plus Auto-contouring, MVision Segmentation Service and RayStation","Radiotherapy centers in the studies reviewed by NICE; NHS radiotherapy services in England; Multiple countries (studies); England, United Kingdom (guidance)",controlled comparison,"79 studies, with data extracted from 15 priority studies, 8 prospective.",Reported time savings ranged from 3 to 80 minutes,Found by an assessment body,"Persisted, 2026-05-15","Of the 79 studies NICE found, 52 were conference abstracts. Some reported time savings left out review and editing time. Only four technologies had evidence on radiation dose, and all lacked long-term patient outcomes such as radiation toxicity. No evidence covered children or people with atypical anatomy. In a 2026 evaluation in eight NHS departments, faster contouring did not shorten the whole planning pathway.",https://petervartanian.xyz/aibd/benefit/AIBD-0654.html
|
| 247 |
AIBD-1669,AI tools in 28 Dhaka pharmacies cut symptom-based antibiotic sales,"United International University's AIMS Lab, with CMED Health Limited",AI-supported antibiotic stewardship system for community pharmacies,"United International University's AIMS Lab, with CMED Health Limited; Dhaka, Bangladesh",before and after,not checked,6.42%,The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-1669.html
|
| 248 |
AIBD-0641,A phone camera coach improved blood pressure in hypertension,RaphAI Health Technologies Limited,Rehabilitation Exercise Prescription app (AI pose recognition built on TensorFlow and OpenPose),"Zhujiang Hospital, Southern Medical University, and Guangdong Provincial People's Hospital; Guangzhou, Guangdong, China",randomized trial,"62 adults aged 18 to 75 with hypertension, Guangzhou.",The app group gained 62.77 meters more in the 6-minute walk test than the control group,Independently studied,not in the sources,"The trial enrolled 62 patients by convenience sampling at two Guangzhou hospitals, was open-label and lasted eight weeks. The authors say that the small sample from one region limits generalizability and that eight weeks cannot show long-term effects.",https://petervartanian.xyz/aibd/benefit/AIBD-0641.html
|
| 249 |
AIBD-1066,EC01-01 helped endoscopists find more colon polyps in a trial,"Xiamen Yingnuo Medical Technology Co., Ltd. (厦门影诺医疗科技有限公司)",EC01-01 (肠息肉电子下消化道图像辅助检测软件),"Xiamen Yingnuo Medical Technology (prospective parallel-controlled trial with 1,020 subjects at four sites); Four trial sites in China that the report does not name",controlled comparison,not checked,"detected 94.7% of polyps, against 84.5% without it",Found by a regulator,"Expanded, 2025-05-22","The trial was the maker's own, a prospective parallel-controlled study at four Chinese sites, and the report does not say whether patients were randomized. The reference came from two experts reading the recorded videos, with a third settling disagreements. CMDE says the output is for reference only and no clinical decision may rest on it alone.",https://petervartanian.xyz/aibd/benefit/AIBD-1066.html
|
|
|
|
| 278 |
AIBD-0770,Samsung's Sleep Apnea Feature caught most moderate to severe cases,"Samsung Electronics Co., Ltd.",Sleep Apnea Feature (Samsung Galaxy Watch and Phone),Samsung (multi-center clinical validation against polysomnography run by an accredited sleep lab); Sites not named in the decision summary,tested against a standard,not checked,sensitivity was 82.7%,Found by a regulator,"Expanded, 2025-06-05","The study enrolled an enriched population, and 47 of 620 enrolled subjects did not finish. The feature gave no usable result on 16.7% of nights. The specificity target was met only after a post-hoc analysis. Subgroups were not powered for conclusive comparisons. FDA says that 10 people were directed to care but reports no follow-up.",https://petervartanian.xyz/aibd/benefit/AIBD-0770.html
|
| 279 |
AIBD-2263,Machine learning sepsis alert speeds antibiotics in a Bangkok trial,Not stated in the abstract. The authors work at Chulalongkorn University.,Sepsis alert by a machine learning model on a staff display (King Chulalongkorn Memorial Hospital),"Emergency department of King Chulalongkorn Memorial Hospital, Bangkok; Bangkok, Thailand",randomized trial,not checked,68.4%,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2263.html
|
| 280 |
AIBD-1065,Dr.Wise-AADS helped doctors find more brain aneurysms on CT,Hangzhou Deepwise (杭州深睿博联科技有限公司),Dr.Wise-AADS (颅内动脉瘤CT造影图像辅助检测软件),"Hangzhou Deepwise (retrospective crossover reader trial on 965 cases at four institutions); Four trial institutions that the report does not name, for a registration in China",controlled comparison,not checked,lesion-level sensitivity for brain aneurysms on CT angiograms 10.85% higher,Found by a regulator,not in the sources,"The trial was the applicant's own retrospective crossover study. Doctors read stored scans, and imaging experts' readings were the reference. It measured sensitivity and specificity, not patient outcomes. The software assists detection of only some aneurysms, those of 3 mm or more.",https://petervartanian.xyz/aibd/benefit/AIBD-1065.html
|
| 281 |
+
AIBD-1161,EndoBRAIN-EYE wins a Japanese insurance add-on for better colonoscopy,"Cybernet Systems Co., Ltd. (サイバネットシステム株式会社), with Showa University Northern Yokohama Hospital and Nagoya University",EndoBRAIN-EYE (内視鏡画像診断支援プログラム),Japanese hospitals and clinics where colonoscopy specialists use it during polyp removal (not named); Japan,before and after,not checked,tumor detection rate from 21.6% to 31.2%,Found by an assessment body,not in the sources,"The 21.6% to 31.2% figure for EndoBRAIN-EYE comes from a trade journal's report of Cybernet's press conference, where it is the result of a clinical performance test. The ministry's Q&A names the product and approval number and cites no figures. The measure is specialists' detection rate, not patient outcomes. The add-on requires a physician with at least five years of experience.",https://petervartanian.xyz/aibd/benefit/AIBD-1161.html
|
| 282 |
AIBD-0742,Wales supports CADe colonoscopy after it found more adenomas in trials,"Medtronic (GI Genius), Olympus (ENDO-AID), Fujifilm (CAD EYE), Wision AI (EndoScreener), Wuhan EndoAngel Medical Technology Company (EndoAngel) and other makers","Computer-aided detection (CADe) colonoscopy systems in the trials that HTW pooled, including GI Genius, ENDO-AID, CAD EYE, EndoScreener and EndoAngel","Endoscopy services in the pooled randomized trials; Multiple countries (trials), with guidance for Wales, United Kingdom",randomized trial,not checked,23% higher with CADe than with standard colonoscopy,Found by an assessment body,not in the sources,"The main source of bias was that endoscopists and outcome assessors were not blinded. HTW reports high heterogeneity and wide confidence intervals. It found no evidence that CADe raised detection of advanced adenomas or of carcinoma, which was rare. No study reported subsequent surgery, time to diagnosis or quality of life. HTW asked for data on real-world use.",https://petervartanian.xyz/aibd/benefit/AIBD-0742.html
|
| 283 |
AIBD-2284,DAX Copilot trims note time for University of Pennsylvania clinicians,Nuance Communications,DAX Copilot,"University of Pennsylvania Health System (outpatient practices); Philadelphia, Pennsylvania, United States",before and after,not checked,from 10.3 to 8.2 minutes,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2284.html
|
| 284 |
AIBD-0626,Sepsis ImmunoScore sorted patients by sepsis risk in FDA's review,"Prenosis, Inc.",Sepsis ImmunoScore,"Prenosis (retrospective validation on data from three US hospitals); Boston, Massachusetts; Chicago, Illinois; Royal Oak, Michigan, United States",tested against a standard,746 patients with suspected sepsis at three US hospitals.,an area under the ROC curve of 0.81,Found by a regulator,"Expanded, 2026-07-28","The study was retrospective, using data collected prospectively at three hospitals, so the score guided no care. Physicians at the patients' own hospitals judged who had sepsis, though FDA recommends independent adjudicators at separate institutions. Only 33 patients fell in the very high group, and its 95% interval for sepsis ran from 51.29% to 84.41%.",https://petervartanian.xyz/aibd/benefit/AIBD-0626.html
|
|
|
|
| 394 |
AIBD-0402,A university plans software that corrects rehabilitation exercises,Anna University,AI-driven exercise assessment and live-feedback system,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"A planned evaluation cannot show that the feedback is accurate or that it improves recovery beyond usual rehabilitation. Safety, adherence and performance in unsupervised homes remain untested.",https://petervartanian.xyz/aibd/benefit/AIBD-0402.html
|
| 395 |
AIBD-0404,A Nigerian team plans software to make impaired speech clearer,AI4Africa,AI4Africa AI-powered AAC tool,Nigeria,not checked,not checked,not in the sources,The makers' claim,not in the sources,The team names Limi Hospital as one of the hospitals that it plans to approach to test the tool. The hospital was not a confirmed testbed. The team has not reported that any person with a speech disability used the tool. Gains in speech intelligibility and communication accuracy require completed development and an evaluation with participants.,https://petervartanian.xyz/aibd/benefit/AIBD-0404.html
|
| 396 |
AIBD-0405,A Tanzanian university plans wearable alerts for abnormal vital signs,Mbeya University of Science and Technology,AI-Based Health Monitoring and Advisory System,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"The team names no pilot, and fewer complications of disease are described as an expected impact. The team gives no detection performance, alert acted on, clinical outcome or comparison of access. Planned biosensors and deeper models further limit what can be said about the present system.",https://petervartanian.xyz/aibd/benefit/AIBD-0405.html
|
| 397 |
+
AIBD-0406,EquiDermAI is meant to reduce gaps between skin tones in AI diagnosis,"Indian Institute of Technology, Madras",EquiDermAI,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"Data augmentation and on-device design do not establish reduced diagnostic disparities. Performance across skin tones, clinical validation and measured effects on access and energy use are still needed.",https://petervartanian.xyz/aibd/benefit/AIBD-0406.html
|
| 398 |
AIBD-0407,A journaling tool in development would give students emotional feedback,Vellore Institute of Technology,StudentSphere AI journaling system,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"In its summary, the team says that an internal pilot is planned, while in its description it calls the product actively deployed. Clinical validation and oversight by psychiatrists remain future work. The figures for accuracy in classifying emotions and for response time are targets, and no completed crisis response or measured fall in stress has been reported.",https://petervartanian.xyz/aibd/benefit/AIBD-0407.html
|
| 399 |
AIBD-0408,Researchers plan hospital tests of a wheelchair steered by brain waves,Vellore Institute of Technology,Brain-wave driven wheelchair for paralysis patients using EEG and ML,not in the sources,no comparison,not checked,not in the sources,The makers' claim,not in the sources,"Training data from ten people do not establish independent control by people with paralysis. The researchers give no participant characteristics, held-out testing, command errors, safety failures, control comparators or sustained mobility outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0408.html
|
| 400 |
AIBD-0409,A planned smartwatch system would monitor facial paralysis,"Vellore Institute of Technology, Chennai",Smart Watch based Integration System for Habituating Face Paralysis Monitoring,not in the sources,not checked,not checked,not in the sources,The makers' claim,not in the sources,"The team labels deployment active while it places mimicry-recognition models, their training data and a suitable watch in future work. No patient result resolves that discrepancy, and earlier intervention and better quality of life are prospective claims.",https://petervartanian.xyz/aibd/benefit/AIBD-0409.html
|
|
|
|
| 413 |
AIBD-0656,NICE lets NHS talking therapies use the Limbic Access chatbot,Limbic,Limbic Access (AI self-referral chatbot),"NHS Talking Therapies for anxiety and depression services in England; England, United Kingdom",controlled comparison,"129,400 patients in NHS Talking Therapies services in England.","referrals rise by 15%, against 6% in control services",Found by an assessment body,Persisted,The 15% rise comes from an observational study. NICE says that most evidence came from the companies or from authors affiliated with them. It found a lack of evidence on people who could not access the service. It says that it is unclear whether the evidence applies to all NHS providers. It says that evidence on time saved is limited.,https://petervartanian.xyz/aibd/benefit/AIBD-0656.html
|
| 414 |
AIBD-0740,NICE allows the Wysa referral chatbot in NHS talking therapies,Wysa,Wysa Digital Referral Assistant (Wysa DRA),"NHS Talking Therapies for anxiety and depression services in England; England, United Kingdom",no comparison,not checked,16.0 to 21.0 minutes per assessment,Independently reported,not in the sources,"NICE's committee had only company-reported, non-comparative data for the Wysa chatbot, none from published research. It could not pool the evidence because of its heterogeneity. It found that evidence on time saved was limited and that the net time saved remains uncertain. It found the evidence's generalisability across NHS Talking Therapies services unclear and the impact on clinical decision making unclear.",https://petervartanian.xyz/aibd/benefit/AIBD-0740.html
|
| 415 |
AIBD-0633,ArteraAI Prostate sorted prostate cancer patients by risk of spread,"Artera, Inc.",ArteraAI Prostate,"Artera (clinical validation at three US sites); Three sites, United States",tested against a standard,886 patients at three US sites.,10-year metastasis risk was 28.1% in its high risk category,Found by a regulator,not in the sources,"The FDA review reports how well the categories separated 10-year risk in 886 patients. It does not show that treatment guided by the software changed outcomes, and FDA says only that the device may help reduce under- or over-treatment. Only 72 patients were African American. The submission held no information on patient perspectives.",https://petervartanian.xyz/aibd/benefit/AIBD-0633.html
|
| 416 |
+
AIBD-0064,"After a scan aided by Brainomix finds a stroke, she leaves in 25 minutes",Brainomix,Brainomix 360 Stroke,"NHS England / Royal Berkshire Hospital; Reading, United Kingdom",no comparison,"109 patients transferred for thrombectomy from five low-volume primary stroke centers in Poland over 53 months, and 69 patients at a comprehensive stroke center.",transferred in 25 minutes for successful thrombectomy,The deployer's account,"Expanded, 2025-12-02",A transfer within 25 minutes and a good individual recovery do not isolate the contribution of Brainomix 360 Stroke. The 64-minute reduction that NHS England reports comes from a different comparison and does not measure the time saved for this patient. NICE stated in January 2024 that it is unclear how much of the faster access to treatment comes from the software.,https://petervartanian.xyz/aibd/benefit/AIBD-0064.html
|
| 417 |
AIBD-0305,ChatGPT lists a rare condition that a genetic test confirms,OpenAI,ChatGPT,"Cardiff, United Kingdom",no comparison,not checked,not in the sources,Independently reported,"Persisted, 2026-10-05","The evidence is the patient's account through two outlets. The reporters do not say that they inspected clinical records, the prompt, the model version or the full list of candidates. Diagnosis alone does not establish a treatment benefit, and reports of how long the coma lasted disagree.",https://petervartanian.xyz/aibd/benefit/AIBD-0305.html
|
| 418 |
AIBD-0134,"ChatGPT flags an insulinoma that doctors missed, which surgery removes",OpenAI,ChatGPT,"Majadahonda, Spain",no comparison,not checked,surgeons removed about two thirds of his pancreas and his symptoms went,Independently reported,"Persisted, 2025-08-07","Every report traces to the account of the patient or his partner, and none gives pathology results or testimony from a clinician who treated him. The only physician quoted, David Céspedes, had no part in his care. Some headlines spoke of pancreatic cancer, but an insulinoma is not by itself evidence of a malignant cancer.",https://petervartanian.xyz/aibd/benefit/AIBD-0134.html
|
| 419 |
AIBD-0136,Staff agree with ChatGPT that a worsening spider bite is an emergency,OpenAI,ChatGPT,"Wolfforth, United States",no comparison,not checked,not in the sources,Independently reported,not in the sources,"No confirmed diagnosis, treatment or follow-up is reported, and the clinicians' comments come only through the woman's account. This case does not show whether ChatGPT saved her life or whether its advice on where to seek care is accurate in general.",https://petervartanian.xyz/aibd/benefit/AIBD-0136.html
|
| 420 |
AIBD-0271,GPT-5 lists oncocytoma first in two retrospective cases,OpenAI,"ChatGPT (GPT-5, 2025 version)","Department of Medicine, Surgery and Pharmacy, University of Sassari; Sassari, Italy",no comparison,"Two patients with caruncular oncocytoma confirmed by histology at the University of Sassari eye clinic, plus three published images of other caruncular tumors.",not in the sources,The deployer's account,not in the sources,"The original two diagnoses were known before testing. Additional tumor tests expose misclassification, but the small selected sample does not establish sensitivity or specificity. Queries were not repeated, and the study did not compare independent clinician assessments or measure prospective patient benefit.",https://petervartanian.xyz/aibd/benefit/AIBD-0271.html
|
| 421 |
AIBD-2178,CHARM gives South African health workers care advice for each client,"Causal Foundry, with mothers2mothers",CHARM,"mothers2mothers; Western Cape, South Africa",no comparison,not checked,not in the sources,The deployer's account,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2178.html
|
| 422 |
AIBD-2176,"Ask-A-Question answers 880,000 questions from South African mothers",IDinsight and Reach Digital Health,Ask-A-Question (AAQ),"National Department of Health of South Africa (MomConnect), with Reach Digital Health; South Africa",no comparison,not checked,"more than 880,000 questions",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2176.html
|
| 423 |
+
AIBD-1091,MIAS-3000 flagged macular degeneration on OCT scans in China's review,"Suzhou Bigvision Medical Technology Co., Ltd. (苏州比格威医疗科技有限公司)",MIAS-3000 (眼科OCT图像辅助诊断软件),Suzhou Bigvision Medical Technology (prospective multicenter registration trial with 908 subjects); Trial sites in China that the report does not name,tested against a standard,not checked,95.5% sensitivity and 92.1% specificity per eye,Found by a regulator,not in the sources,"The trial was the maker's own, a prospective multicenter single-group study of 908 subjects in which doctors read with the software in one group and alone in the other. CMDE reports sensitivity and specificity per eye and says that clinical decisions must not rest on the product's result alone. No change in any patient's care is reported.",https://petervartanian.xyz/aibd/benefit/AIBD-1091.html
|
| 424 |
AIBD-0739,AI tools find spinal fractures missed in routine scan reports,"Aidoc Medical (BriefCase-Triage), Avicenna.AI (CINA-VCF Quantix), Nanox AI (HealthVCF and HealthOST) and IB Lab (IB Lab FLAMINGO)","AI tools for opportunistic detection of vertebral fragility fractures on CT: BriefCase-Triage, CINA-VCF Quantix, HealthVCF, HealthOST and IB Lab FLAMINGO","Hospitals in the studies reviewed by NICE and NHS radiology services in England during the evidence generation period; England, United Kingdom (guidance), with studies in several countries",tested against a standard,not checked,a specificity above 90%,Found by an assessment body,Persisted,"Most studies of the AI tools were retrospective and measured diagnostic accuracy, and most comparisons with standard care were done outside the UK. NICE says that it is uncertain how much the tools improve detection in the NHS. Very little evidence shows how they affect clinical management. The model's standard-care values came from a small expert elicitation study.",https://petervartanian.xyz/aibd/benefit/AIBD-0739.html
|
| 425 |
AIBD-0959,AI-Rad Companion Prostate MR lifted readers' accuracy slightly,"Siemens Healthineers, with 上海西门子医疗器械有限公司 (Siemens Shanghai Medical Equipment Ltd.) as the applicant in China",AI-Rad Companion Prostate MR (前列腺癌磁共振图像辅助检测软件),"上海西门子医疗器械有限公司 (Siemens Shanghai Medical Equipment Ltd.), the applicant (retrospective reader trial with 12 doctors and 340 subjects); Trial sites that the report does not name, for a registration in China",controlled comparison,not checked,"an AUROC of 0.7010, against 0.6758 without it",Found by a regulator,"Expanded, 2025-11-07","The trial was retrospective, with 12 doctors reading images from 340 subjects. Siemens Shanghai, the applicant, supplied the trial data. The AUROC gain of 0.0252 had a 95% confidence interval from 0.0011 to 0.0493. The sensitivity gain of 4.4% had an interval from -0.2% to 9.1%. Median reading time was 98 seconds in both groups.",https://petervartanian.xyz/aibd/benefit/AIBD-0959.html
|
| 426 |
AIBD-2185,AMTI answers refugee women's health questions on WhatsApp in Türkiye,"MEDAK, with HERA Digital Health",AMTI (Predictive Care Navigator),"MEDAK, with HERA Digital Health; Southeast Türkiye",no comparison,not checked,"nearly 2,500 users",The makers' claim,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2185.html
|
|
|
|
| 434 |
AIBD-0912,Norfolk's AI tool for preventing falls is credited with fewer admissions,Not named in the assessment,AI-driven proactive falls prevention tool,"Norfolk County Council, adult social care; Norfolk, England, United Kingdom",no comparison,not checked,not in the sources,Independently reported,"Expanded, 2025-12-12",The evidence is council leaders' account of a pilot and partner feedback. CQC says that data supported it but gives no figures and does not say whose data it saw.,https://petervartanian.xyz/aibd/benefit/AIBD-0912.html
|
| 435 |
AIBD-1033,AI-based image analysis detects referable diabetic retinopathy,"Several vendors, among them Eyenuk (EyeArt), Digital Diagnostics (IDx-DR) and EyRIS (SELENA)","AI-based image analysis of fundus photographs for diabetic retinopathy screening (products in the included studies such as EyeArt, IDx-DR, SELENA, RetCAD and Aireen)","Clinics and screening programs in the comparative studies that HTA syd reviewed; Studies in Europe, Asia, Africa, Oceania and North America, assessed for Sweden's Southern Healthcare Region",tested against a standard,not checked,about 92% sensitivity and 91% specificity,Found by an assessment body,"Persisted, 2026-04-22","HTA syd's pooled accuracy comes from 13 non-randomized studies that compared AI grading with human graders, all from other health systems. HTA syd could not judge whether recalls, future vision loss or blindness change with AI grading. Sweden's grading scale differs from the international standard, which makes the results hard to transfer. Moderate certainty means that more research is needed, and no current cost-effectiveness analysis exists.",https://petervartanian.xyz/aibd/benefit/AIBD-1033.html
|
| 436 |
AIBD-0023,An Apple Watch called for help after a runner’s heart stopped,Apple,Apple Watch Fall Detection,"Varsseveld, Gelderland, Netherlands",no comparison,not checked,began cardiopulmonary resuscitation about 6–7 minutes after the collapse,Independently studied,"Persisted, 2026-05-01","The authors give no date or place. The identification of the patient, the date and the street are inferred from interviews and a pager message. Men's Health says that the runner also pressed the SOS button himself and that a pulse returned within about three minutes. This account does not fit the timeline in the paper. The case cannot show how quickly he would have been found without the alert.",https://petervartanian.xyz/aibd/benefit/AIBD-0023.html
|
| 437 |
+
AIBD-0891,Clinicians feel less burdened by documentation with AI scribes,"Several makers, including Nuance (Dragon Ambient eXperience), Abridge, Nabla, Sunoh.ai and Amazon","AI scribes (ambient scribes), including Dragon Ambient eXperience, Abridge, Nabla Copilot, Sunoh.ai and Amazon HealthScribe","Hospitals, clinics and other services in the studies that AIHTA reviewed; Multiple countries (studies), mostly the United States",no comparison,not checked,from a 19-92% decrease to a 13-50% increase,Found by an assessment body,"Persisted, 2026-05-01","AIHTA pools seven reviews of many AI scribe products, so no single system is assessed. Satisfaction and burden are clinicians' own reports, and measured documentation time ranged from large decreases to increases. Örebro's review of the four randomized trials found self-selected participants, undefined control conditions and high or unacceptably high risk of bias in every trial.",https://petervartanian.xyz/aibd/benefit/AIBD-0891.html
|
| 438 |
AIBD-2276,AI scribe adopters at UCSF Health bill more work units per week,Not named in the report,Ambient AI scribes (products not named in the report),"UCSF Health (ambulatory practices); San Francisco, California, United States",controlled comparison,not checked,1.81 relative value units per week,Independently studied,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2276.html
|
| 439 |
AIBD-0156,Ward staff revive a patient after thynC flags a pre-arrest heart rhythm,"Seers Technology Co., Seongnam (maker of the ThynC patch); Daewoong Pharmaceutical supplies thynC","thynC (씽크) AI-based smart bed monitoring system: wearable sensors measuring ECG, heart rate and respiratory rate in real time, with alarms sent to a remote central monitor","Gwangju Hanguk Hospital (광주한국병원); Gwangju, South Korea",no comparison,not checked,heart rate climbed to 192 beats per minute,The makers' claim,"Expanded, 2026-06-19","The account comes through the supplier, with no clinical record and no response time that anyone independent has assessed. Staff resuscitated the patient, and the time that would have passed without the alert is unknown. A restored circulation leaves long-term recovery unknown.",https://petervartanian.xyz/aibd/benefit/AIBD-0156.html
|
| 440 |
AIBD-0267,Mother's AI query prompts a rare-disease diagnosis in a 22-month-old,Google (Med-Gemini-3D),Smartphone application described by the authors as powered by Med-Gemini-3D,"The patient's mother (consumer use); United States (authors at Garnet Health Medical Center, Middletown, New York)",no comparison,not checked,not in the sources,Independently studied,not in the sources,"In their abstract, the authors do not identify the app or show that it ran Med-Gemini-3D. They do not say which therapeutic interventions followed or what they achieved.",https://petervartanian.xyz/aibd/benefit/AIBD-0267.html
|
|
|
|
| 475 |
AIBD-0274,Lunit INSIGHT CXR flags a subpulmonic effusion that ultrasound confirms,"Lunit Inc., Seoul",Lunit INSIGHT CXR,"Basic Emergency Service, Vila Real de Santo Antonio Health Center, Algarve Local Health Unit; Vila Real de Santo Antonio, Algarve, Portugal",no comparison,not checked,not in the sources,The deployer's account,not in the sources,"Correct identification does not establish subsequent clinical improvement. The patient's refusal limits the observed outcome, and the retained case does not isolate what the radiographer would have recognized without the AI output.",https://petervartanian.xyz/aibd/benefit/AIBD-0274.html
|
| 476 |
AIBD-2277,Ambient AI scribe shortens note time in an emergency department,Nuance Communications,"Ambient AI scribe (DAX Copilot, as the outside review names it)","A tertiary academic medical center in California (the authors work at Stanford); Stanford, California, United States",controlled comparison,not checked,2 minutes 45 seconds,Independently studied,"Persisted, 2026-07-02",not checked,https://petervartanian.xyz/aibd/benefit/AIBD-2277.html
|
| 477 |
AIBD-0957,uAI-BreastTomoCare helped doctors find more breast masses,"United Imaging Intelligence (Beijing) Co., Ltd. (联影智能医疗科技(北京)有限公司)",uAI-BreastTomoCare (乳腺X射线断层图像辅助检测软件),"United Imaging Intelligence (Beijing) (retrospective reader trial on images of 432 subjects from three centers); Three trial centers that the report does not name, for a registration in China",controlled comparison,not checked,"62.17% of breast mass lesions on tomosynthesis, against 50.42% without it",Found by a regulator,not in the sources,"The trial was the maker's own retrospective reading study of stored images from three centers. It measured mass detection, not outcomes for patients. Used alone, the software had a specificity of 53.81% at the patient level. The review says that trained doctors must use it and that it cannot serve alone as the basis for clinical decisions.",https://petervartanian.xyz/aibd/benefit/AIBD-0957.html
|
| 478 |
+
AIBD-0788,Claire OCT System helped surgeons remove cancer left at the margins,"Perimeter Medical Imaging AI, Inc.",Claire OCT System with the ImgAssist AI detection algorithm,"Breast surgeons at the investigational sites of the randomized trial (NCT05113927); Investigational sites, United States",controlled comparison,not checked,a 20.0% relative reduction,Found by a regulator,not in the sources,"The Claire OCT System was used only after surgeons finished standard margin assessment, so its gain is measured as an add-on in 206 device-arm patients, with 7 patients converted to clear margins. FDA notes that real-world use may differ from trial conditions. Margin-level sensitivity and specificity were assessed post hoc, and no outcome such as recurrence is reported.",https://petervartanian.xyz/aibd/benefit/AIBD-0788.html
|
| 479 |
AIBD-0767,Tyto Insights detects bulging eardrums better than non-ENT doctors,Tyto Care Ltd.,Tyto Insights for Eardrum Bulging Detection,Tyto Care (retrospective validation on 857 otoscope recordings from real-world use); Recording sites not stated in the decision summary,tested against a standard,not checked,a sensitivity of 0.9231,Found by a regulator,"Persisted, 2026-04-27",not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0767.html
|
| 480 |
AIBD-0653,England's health technology assessor supports AI help in colonoscopy,"Odin Vision (CADDIE), Fujifilm (CAD EYE), Olympus Medical Systems (ENDO-AID), Wision AI (EndoScreener), Medtronic (GI Genius) and Magentiq Eye (MAGENTIQ-COLO)","AI computer-aided detection (CADe) for colonoscopy: CADDIE, CAD EYE, ENDO-AID, EndoScreener, GI Genius and MAGENTIQ-COLO","Endoscopy services in the trials reviewed by NICE; NHS services in England during the evidence generation period; Multiple countries (trials); England, United Kingdom (guidance)",randomized trial,"70 studies of AI-assisted live colonoscopy, mostly randomized controlled trials.",not in the sources,Found by an assessment body,not in the sources,"NICE's committee judged it uncertain whether the higher adenoma detection rate will lower colorectal cancer rates. Gains for advanced adenomas and sessile serrated lesions were not statistically significant for any system. Most studies excluded people with Lynch syndrome, familial adenomatous polyposis or inflammatory bowel disease. Data on false positives were very limited. The quality and quantity of evidence varied across systems.",https://petervartanian.xyz/aibd/benefit/AIBD-0653.html
|
| 481 |
AIBD-0741,ArtiQ.Spiro raised sensitivity for COPD in a UK trial,Clario,ArtiQ.Spiro,SPIRO-AID trial with 133 UK primary care professionals reviewing past spirometry records; United Kingdom,randomized trial,not checked,not in the sources,Found by an assessment body,not in the sources,"In SPIRO-AID, professionals reviewed 50 past spirometry tests, a small sample with few normal results. The external group calculated accuracy from data that the authors supplied in confidence. Accuracy in primary care use is unclear, and no study reported long-term outcomes.",https://petervartanian.xyz/aibd/benefit/AIBD-0741.html
|
for-procurement.csv
CHANGED
|
@@ -1456,7 +1456,7 @@ AIBD-0401,A university is developing a moral-tale game for autistic children,Chr
|
|
| 1456 |
AIBD-0402,A university plans software that corrects rehabilitation exercises,Anna University,AI-driven exercise assessment and live-feedback system,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0402.html
|
| 1457 |
AIBD-0404,A Nigerian team plans software to make impaired speech clearer,AI4Africa,AI4Africa AI-powered AAC tool,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,Clearer speech would help only if the output preserves what the person meant. No such error in this tool has been reported.,https://petervartanian.xyz/aibd/benefit/AIBD-0404.html
|
| 1458 |
AIBD-0405,A Tanzanian university plans wearable alerts for abnormal vital signs,Mbeya University of Science and Technology,AI-Based Health Monitoring and Advisory System,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not in the sources,https://petervartanian.xyz/aibd/benefit/AIBD-0405.html
|
| 1459 |
-
AIBD-0406,EquiDermAI is meant to reduce gaps between skin tones in AI diagnosis,
|
| 1460 |
AIBD-0407,A journaling tool in development would give students emotional feedback,Vellore Institute of Technology,StudentSphere AI journaling system,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,"Costs and harms remain unassessed, and no change in well-being or escalation to a counselor has been reported for any student. Nothing shows that the same harm occurred here.",https://petervartanian.xyz/aibd/benefit/AIBD-0407.html
|
| 1461 |
AIBD-0408,Researchers plan hospital tests of a wheelchair steered by brain waves,Vellore Institute of Technology,Brain-wave driven wheelchair for paralysis patients using EEG and ML,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0408.html
|
| 1462 |
AIBD-0409,A planned smartwatch system would monitor facial paralysis,"Vellore Institute of Technology, Chennai",Smart Watch based Integration System for Habituating Face Paralysis Monitoring,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not in the sources,https://petervartanian.xyz/aibd/benefit/AIBD-0409.html
|
|
|
|
| 1456 |
AIBD-0402,A university plans software that corrects rehabilitation exercises,Anna University,AI-driven exercise assessment and live-feedback system,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0402.html
|
| 1457 |
AIBD-0404,A Nigerian team plans software to make impaired speech clearer,AI4Africa,AI4Africa AI-powered AAC tool,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,Clearer speech would help only if the output preserves what the person meant. No such error in this tool has been reported.,https://petervartanian.xyz/aibd/benefit/AIBD-0404.html
|
| 1458 |
AIBD-0405,A Tanzanian university plans wearable alerts for abnormal vital signs,Mbeya University of Science and Technology,AI-Based Health Monitoring and Advisory System,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not in the sources,https://petervartanian.xyz/aibd/benefit/AIBD-0405.html
|
| 1459 |
+
AIBD-0406,EquiDermAI is meant to reduce gaps between skin tones in AI diagnosis,"Indian Institute of Technology, Madras",EquiDermAI,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,Editor's note: Compact local models reduce reliance on a remote service. The fairness objective depends on synthetic images representing real lesions accurately rather than merely balancing image counts.,https://petervartanian.xyz/aibd/benefit/AIBD-0406.html
|
| 1460 |
AIBD-0407,A journaling tool in development would give students emotional feedback,Vellore Institute of Technology,StudentSphere AI journaling system,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,"Costs and harms remain unassessed, and no change in well-being or escalation to a counselor has been reported for any student. Nothing shows that the same harm occurred here.",https://petervartanian.xyz/aibd/benefit/AIBD-0407.html
|
| 1461 |
AIBD-0408,Researchers plan hospital tests of a wheelchair steered by brain waves,Vellore Institute of Technology,Brain-wave driven wheelchair for paralysis patients using EEG and ML,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not checked,https://petervartanian.xyz/aibd/benefit/AIBD-0408.html
|
| 1462 |
AIBD-0409,A planned smartwatch system would monitor facial paralysis,"Vellore Institute of Technology, Chennai",Smart Watch based Integration System for Habituating Face Paralysis Monitoring,not in the sources,The makers' claim,not in the sources,not in the sources,not in the sources,not in the sources,not in the sources,https://petervartanian.xyz/aibd/benefit/AIBD-0409.html
|
manifest.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
profiles.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:140a113e87682473fdbc11cadbf43ef8b389bfe7658d36041bd14675cbb666ac
|
| 3 |
+
size 63849730
|
querist.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
querist.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf2b452db84d557e4c8574c296d85a084ebd9e93b2ed9a593f14bd8f3dd66c68
|
| 3 |
+
size 21817151
|
records.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
records.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:28947ea2eada0aff91ff3dd214748abf20afc8beba658df53bd590ebe33794c6
|
| 3 |
+
size 80674146
|
ro-crate-metadata.json
CHANGED
|
@@ -148,8 +148,8 @@
|
|
| 148 |
"name": "adyt.csv",
|
| 149 |
"description": "The benefits of the collection Adyt.",
|
| 150 |
"encodingFormat": "text/csv",
|
| 151 |
-
"contentSize": "
|
| 152 |
-
"sha256": "
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"@id": "adyt.json",
|
|
@@ -157,8 +157,8 @@
|
|
| 157 |
"name": "adyt.json",
|
| 158 |
"description": "The benefits of the collection Adyt.",
|
| 159 |
"encodingFormat": "application/json",
|
| 160 |
-
"contentSize": "
|
| 161 |
-
"sha256": "
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"@id": "chanticleer.csv",
|
|
@@ -166,8 +166,8 @@
|
|
| 166 |
"name": "chanticleer.csv",
|
| 167 |
"description": "The benefits of the collection Chanticleer.",
|
| 168 |
"encodingFormat": "text/csv",
|
| 169 |
-
"contentSize": "
|
| 170 |
-
"sha256": "
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"@id": "chanticleer.json",
|
|
@@ -175,8 +175,8 @@
|
|
| 175 |
"name": "chanticleer.json",
|
| 176 |
"description": "The benefits of the collection Chanticleer.",
|
| 177 |
"encodingFormat": "application/json",
|
| 178 |
-
"contentSize": "
|
| 179 |
-
"sha256": "
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"@id": "collections.json",
|
|
@@ -202,8 +202,8 @@
|
|
| 202 |
"name": "for-clinical-use.csv",
|
| 203 |
"description": "The benefits under the questions of one document: The Joint Commission and the Coalition for Health AI, Guidance on the Responsible Use of AI in Healthcare (RUAIH), Element 4, Ongoing Quality Monitoring.",
|
| 204 |
"encodingFormat": "text/csv",
|
| 205 |
-
"contentSize": "
|
| 206 |
-
"sha256": "
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"@id": "for-court-use.csv",
|
|
@@ -229,8 +229,8 @@
|
|
| 229 |
"name": "for-procurement.csv",
|
| 230 |
"description": "The benefits under the questions of one document: OMB Memorandum M-25-22, Driving Efficient Acquisition of Artificial Intelligence in Government, Section 4(b)(iii), Performance-Based Acquisition Techniques.",
|
| 231 |
"encodingFormat": "text/csv",
|
| 232 |
-
"contentSize": "
|
| 233 |
-
"sha256": "
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"@id": "for-regulation.csv",
|
|
@@ -301,8 +301,8 @@
|
|
| 301 |
"name": "profiles.json",
|
| 302 |
"description": "The public profile of each benefit.",
|
| 303 |
"encodingFormat": "application/json",
|
| 304 |
-
"contentSize": "
|
| 305 |
-
"sha256": "
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"@id": "querist.csv",
|
|
@@ -310,8 +310,8 @@
|
|
| 310 |
"name": "querist.csv",
|
| 311 |
"description": "The benefits of the collection Querist.",
|
| 312 |
"encodingFormat": "text/csv",
|
| 313 |
-
"contentSize": "
|
| 314 |
-
"sha256": "
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"@id": "querist.json",
|
|
@@ -319,8 +319,8 @@
|
|
| 319 |
"name": "querist.json",
|
| 320 |
"description": "The benefits of the collection Querist.",
|
| 321 |
"encodingFormat": "application/json",
|
| 322 |
-
"contentSize": "
|
| 323 |
-
"sha256": "
|
| 324 |
},
|
| 325 |
{
|
| 326 |
"@id": "ratings.csv",
|
|
@@ -346,8 +346,8 @@
|
|
| 346 |
"name": "records.csv",
|
| 347 |
"description": "One row per benefit.",
|
| 348 |
"encodingFormat": "text/csv",
|
| 349 |
-
"contentSize": "
|
| 350 |
-
"sha256": "
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"@id": "records.json",
|
|
@@ -355,8 +355,8 @@
|
|
| 355 |
"name": "records.json",
|
| 356 |
"description": "The dataset: every benefit with its sources.",
|
| 357 |
"encodingFormat": "application/json",
|
| 358 |
-
"contentSize": "
|
| 359 |
-
"sha256": "
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"@id": "requirements.json",
|
|
@@ -391,8 +391,8 @@
|
|
| 391 |
"name": "viator.csv",
|
| 392 |
"description": "The benefits of the collection Viator.",
|
| 393 |
"encodingFormat": "text/csv",
|
| 394 |
-
"contentSize": "
|
| 395 |
-
"sha256": "
|
| 396 |
},
|
| 397 |
{
|
| 398 |
"@id": "viator.json",
|
|
@@ -400,8 +400,8 @@
|
|
| 400 |
"name": "viator.json",
|
| 401 |
"description": "The benefits of the collection Viator.",
|
| 402 |
"encodingFormat": "application/json",
|
| 403 |
-
"contentSize": "
|
| 404 |
-
"sha256": "
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"@id": "datapackage.json",
|
|
@@ -410,7 +410,7 @@
|
|
| 410 |
"description": "This folder as a Frictionless Data Package.",
|
| 411 |
"encodingFormat": "application/json",
|
| 412 |
"contentSize": "64950",
|
| 413 |
-
"sha256": "
|
| 414 |
"conformsTo": {
|
| 415 |
"@id": "https://datapackage.org/profiles/2.0/datapackage.json"
|
| 416 |
}
|
|
|
|
| 148 |
"name": "adyt.csv",
|
| 149 |
"description": "The benefits of the collection Adyt.",
|
| 150 |
"encodingFormat": "text/csv",
|
| 151 |
+
"contentSize": "370189",
|
| 152 |
+
"sha256": "46b3a378638ed98fa822d9acff6ed8cb63a2f5514d441313e98e6bc7784f3426"
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"@id": "adyt.json",
|
|
|
|
| 157 |
"name": "adyt.json",
|
| 158 |
"description": "The benefits of the collection Adyt.",
|
| 159 |
"encodingFormat": "application/json",
|
| 160 |
+
"contentSize": "9887959",
|
| 161 |
+
"sha256": "3770d07489a41cedbb8eb67643bd99f1c9be462b1056c86cbb5f2c37602c4703"
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"@id": "chanticleer.csv",
|
|
|
|
| 166 |
"name": "chanticleer.csv",
|
| 167 |
"description": "The benefits of the collection Chanticleer.",
|
| 168 |
"encodingFormat": "text/csv",
|
| 169 |
+
"contentSize": "342483",
|
| 170 |
+
"sha256": "302e5597c5f2f886ce71f30e7728ac8774e2b69f1b9bcf2f6bc79816fdfd49d3"
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"@id": "chanticleer.json",
|
|
|
|
| 175 |
"name": "chanticleer.json",
|
| 176 |
"description": "The benefits of the collection Chanticleer.",
|
| 177 |
"encodingFormat": "application/json",
|
| 178 |
+
"contentSize": "10271172",
|
| 179 |
+
"sha256": "89e61355a5bf424cbc24c5acc2c34c08685c2f643ac0e7f04ca3acb51524c986"
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"@id": "collections.json",
|
|
|
|
| 202 |
"name": "for-clinical-use.csv",
|
| 203 |
"description": "The benefits under the questions of one document: The Joint Commission and the Coalition for Health AI, Guidance on the Responsible Use of AI in Healthcare (RUAIH), Element 4, Ongoing Quality Monitoring.",
|
| 204 |
"encodingFormat": "text/csv",
|
| 205 |
+
"contentSize": "330914",
|
| 206 |
+
"sha256": "738d14448666db5ed2d3f84ceb87062338e1bf57ca15a05e4f081be367b4eed3"
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"@id": "for-court-use.csv",
|
|
|
|
| 229 |
"name": "for-procurement.csv",
|
| 230 |
"description": "The benefits under the questions of one document: OMB Memorandum M-25-22, Driving Efficient Acquisition of Artificial Intelligence in Government, Section 4(b)(iii), Performance-Based Acquisition Techniques.",
|
| 231 |
"encodingFormat": "text/csv",
|
| 232 |
+
"contentSize": "1188341",
|
| 233 |
+
"sha256": "85e85c93a3d5ac105fd1e57c8e444eb71bb5dad43992c5ca50a7ad541a31e681"
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"@id": "for-regulation.csv",
|
|
|
|
| 301 |
"name": "profiles.json",
|
| 302 |
"description": "The public profile of each benefit.",
|
| 303 |
"encodingFormat": "application/json",
|
| 304 |
+
"contentSize": "63849730",
|
| 305 |
+
"sha256": "140a113e87682473fdbc11cadbf43ef8b389bfe7658d36041bd14675cbb666ac"
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"@id": "querist.csv",
|
|
|
|
| 310 |
"name": "querist.csv",
|
| 311 |
"description": "The benefits of the collection Querist.",
|
| 312 |
"encodingFormat": "text/csv",
|
| 313 |
+
"contentSize": "1123052",
|
| 314 |
+
"sha256": "585b45110e770f78e6f37161d575f18584d90c64e954a7f36125426021327c14"
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"@id": "querist.json",
|
|
|
|
| 319 |
"name": "querist.json",
|
| 320 |
"description": "The benefits of the collection Querist.",
|
| 321 |
"encodingFormat": "application/json",
|
| 322 |
+
"contentSize": "21817151",
|
| 323 |
+
"sha256": "bf2b452db84d557e4c8574c296d85a084ebd9e93b2ed9a593f14bd8f3dd66c68"
|
| 324 |
},
|
| 325 |
{
|
| 326 |
"@id": "ratings.csv",
|
|
|
|
| 346 |
"name": "records.csv",
|
| 347 |
"description": "One row per benefit.",
|
| 348 |
"encodingFormat": "text/csv",
|
| 349 |
+
"contentSize": "2387733",
|
| 350 |
+
"sha256": "af795b78586f1e0860c2439c2a8c958554f0f3ea37d12b6a64edb2fecf58769f"
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"@id": "records.json",
|
|
|
|
| 355 |
"name": "records.json",
|
| 356 |
"description": "The dataset: every benefit with its sources.",
|
| 357 |
"encodingFormat": "application/json",
|
| 358 |
+
"contentSize": "80674146",
|
| 359 |
+
"sha256": "28947ea2eada0aff91ff3dd214748abf20afc8beba658df53bd590ebe33794c6"
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"@id": "requirements.json",
|
|
|
|
| 391 |
"name": "viator.csv",
|
| 392 |
"description": "The benefits of the collection Viator.",
|
| 393 |
"encodingFormat": "text/csv",
|
| 394 |
+
"contentSize": "577755",
|
| 395 |
+
"sha256": "2a962a7f13d0cdd0e8018a2c2fddfcc081a2f5dd80a43531f332431d06bc1345"
|
| 396 |
},
|
| 397 |
{
|
| 398 |
"@id": "viator.json",
|
|
|
|
| 400 |
"name": "viator.json",
|
| 401 |
"description": "The benefits of the collection Viator.",
|
| 402 |
"encodingFormat": "application/json",
|
| 403 |
+
"contentSize": "15455095",
|
| 404 |
+
"sha256": "a794d3374a028644f6852374cc52ce0217799efeaf44ac5dc65e30be9791d673"
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"@id": "datapackage.json",
|
|
|
|
| 410 |
"description": "This folder as a Frictionless Data Package.",
|
| 411 |
"encodingFormat": "application/json",
|
| 412 |
"contentSize": "64950",
|
| 413 |
+
"sha256": "b29e43d86e6d5dbc741533bfb2919942a030ef75d1ad8319f7a4ae65d72001b9",
|
| 414 |
"conformsTo": {
|
| 415 |
"@id": "https://datapackage.org/profiles/2.0/datapackage.json"
|
| 416 |
}
|
viator.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viator.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a794d3374a028644f6852374cc52ce0217799efeaf44ac5dc65e30be9791d673
|
| 3 |
+
size 15455095
|