Download train2.csv from Devseis/eu-ai-act-classifier-experiment: direct link, hf CLI and curl.
- Browser
- Download file 35.4 kB
-
https://huggingface.co/Devseis/eu-ai-act-classifier-experiment/resolve/main/train2.csv
- Command line
-
hf download hf://Devseis/eu-ai-act-classifier-experiment/train2.csv
-
curl -L -o train2.csv https://huggingface.co/Devseis/eu-ai-act-classifier-experiment/resolve/main/train2.csv
35.4 kB
| text,label | |
| Proposed use case: an AI system designed to decide which employees get promoted based on automated performance scores.,high_risk | |
| Profile students' personalities alone to predict future criminal behaviour.,prohibited | |
| A public sector agency is considering an AI system to grade student exams and determine whether they pass or fail a course.,high_risk | |
| Internal project brief: build an AI capability to auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| Is it in scope if our AI helps generate a realistic video of a public figure saying something they never said?,limited_risk | |
| The vendor's AI feature would balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| Proposed use case: an AI system designed to build a 'trustworthiness score' for citizens combining unrelated data sources and use it to deny public benefits.,prohibited | |
| A public sector agency is considering an AI system to detect emotional states of call-centre agents to evaluate their performance.,high_risk | |
| Detect duplicate rows in a spreadsheet import.,minimal_risk | |
| Internal project brief: build an AI capability to detect duplicate rows in a spreadsheet import.,minimal_risk | |
| Proposed use case: an AI system designed to detect duplicate rows in a spreadsheet import.,minimal_risk | |
| Our client wants to deploy an AI tool to act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| A public sector agency is considering an AI system to use subliminal or manipulative techniques to distort a person's behaviour in a way that causes them significant harm.,prohibited | |
| A startup is building a product that uses AI to control safety-critical operations of a regional electricity grid.,high_risk | |
| "Our client wants to deploy an AI tool to sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| We're planning to use machine learning to detect emotions of pupils during exams to flag 'suspicious' students.,prohibited | |
| We're planning to use machine learning to balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| Our client wants to deploy an AI tool to detect emotional states of call-centre agents to evaluate their performance.,high_risk | |
| Internal project brief: build an AI capability to profile students' personalities alone to predict future criminal behaviour.,prohibited | |
| A public sector agency is considering an AI system to balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| A public sector agency is considering an AI system to detect duplicate rows in a spreadsheet import.,minimal_risk | |
| We're planning to use machine learning to filter obvious spam emails out of a company inbox.,minimal_risk | |
| Internal project brief: build an AI capability to decide which employees get promoted based on automated performance scores.,high_risk | |
| "A startup is building a product that uses AI to predict whether a specific named individual will commit a crime based only on their personality profile, with no other evidence.",prohibited | |
| Proposed use case: an AI system designed to use subliminal or manipulative techniques to distort a person's behaviour in a way that causes them significant harm.,prohibited | |
| Internal project brief: build an AI capability to categorise people using biometric data to infer demographic or protected attributes for marketing analytics.,high_risk | |
| "A public sector agency is considering an AI system to recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| Is it in scope if our AI helps run an AI avatar that answers HR questions for new employees?,limited_risk | |
| Proposed use case: an AI system designed to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| "Is it in scope if our AI helps sort support tickets into internal categories for an engineering team, with no public-facing interaction?",minimal_risk | |
| Is it in scope if our AI helps detect emotional states of call-centre agents to evaluate their performance?,high_risk | |
| A public sector agency is considering an AI system to mass-scrape images from the open internet to expand a biometric identification database.,prohibited | |
| A startup is building a product that uses AI to exploit the vulnerabilities of elderly or disabled users to influence their purchasing decisions against their interests.,prohibited | |
| Internal project brief: build an AI capability to balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| Is it in scope if our AI helps profile students' personalities alone to predict future criminal behaviour?,prohibited | |
| A startup is building a product that uses AI to identify individuals in security footage after the fact by matching faces against a watchlist.,high_risk | |
| Is it in scope if our AI helps act as a virtual assistant that negotiates prices with customers in a live chat?,limited_risk | |
| Proposed use case: an AI system designed to auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| Is it in scope if our AI helps flag students for cheating during an online proctored exam?,high_risk | |
| Is it in scope if our AI helps determine an individual's eligibility for unemployment or housing benefits?,high_risk | |
| We're planning to use machine learning to scrape facial images from social media and CCTV footage to build a facial recognition database without consent.,prohibited | |
| "Score citizens across unrelated contexts (social media, shopping, health) and deny them services based on that combined score.",prohibited | |
| Determine which reading level a school should place a child in.,high_risk | |
| Proposed use case: an AI system designed to infer employees' emotional states during video meetings to flag who might be unhappy at work.,prohibited | |
| Internal project brief: build an AI capability to determine which reading level a school should place a child in.,high_risk | |
| The vendor's AI feature would control safety-critical operations of a regional electricity grid.,high_risk | |
| The vendor's AI feature would run an AI avatar that answers HR questions for new employees.,limited_risk | |
| A startup is building a product that uses AI to detect emotions of pupils during exams to flag 'suspicious' students.,prohibited | |
| We're planning to use machine learning to generate a realistic video of a public figure saying something they never said.,limited_risk | |
| Is it in scope if our AI helps categorise people using biometric data to infer demographic or protected attributes for marketing analytics?,high_risk | |
| Our client wants to deploy an AI tool to compress and resize product photos automatically before upload.,minimal_risk | |
| We're planning to use machine learning to categorise people using biometric data to infer demographic or protected attributes for marketing analytics.,high_risk | |
| Our client wants to deploy an AI tool to flag students for cheating during an online proctored exam.,high_risk | |
| A public sector agency is considering an AI system to run an AI avatar that answers HR questions for new employees.,limited_risk | |
| "Recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| "A public sector agency is considering an AI system to score citizens across unrelated contexts (social media, shopping, health) and deny them services based on that combined score.",prohibited | |
| A public sector agency is considering an AI system to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| We are developing an AI system that will help a judge search case law and summarise legal precedent for a ruling.,high_risk | |
| The vendor's AI feature would compress and resize product photos automatically before upload.,minimal_risk | |
| We are developing an AI system that will mass-scrape images from the open internet to expand a biometric identification database.,prohibited | |
| Is it in scope if our AI helps compress and resize product photos automatically before upload?,minimal_risk | |
| Internal project brief: build an AI capability to run an AI avatar that answers HR questions for new employees.,limited_risk | |
| Generate a realistic video of a public figure saying something they never said.,limited_risk | |
| "The vendor's AI feature would score citizens across unrelated contexts (social media, shopping, health) and deny them services based on that combined score.",prohibited | |
| Is it in scope if our AI helps grade student exams and determine whether they pass or fail a course?,high_risk | |
| The vendor's AI feature would auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| The vendor's AI feature would predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| Decide which employees get promoted based on automated performance scores.,high_risk | |
| Infer employees' emotional states during video meetings to flag who might be unhappy at work.,prohibited | |
| We are developing an AI system that will set individualised life insurance premiums based on predicted health risk.,high_risk | |
| Proposed use case: an AI system designed to determine an individual's eligibility for unemployment or housing benefits.,high_risk | |
| We're planning to use machine learning to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| Is it in scope if our AI helps swap a person's face into an existing video for a marketing campaign?,limited_risk | |
| The vendor's AI feature would set individualised life insurance premiums based on predicted health risk.,high_risk | |
| We are developing an AI system that will clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Our client wants to deploy an AI tool to write blog posts that are published on our site with no human review.,limited_risk | |
| "We're planning to use machine learning to predict whether a specific named individual will commit a crime based only on their personality profile, with no other evidence.",prohibited | |
| We are developing an AI system that will evaluate the reliability of evidence in an ongoing criminal investigation.,high_risk | |
| Is it in scope if our AI helps auto-generate a news-style article on a current political topic for publication?,limited_risk | |
| A startup is building a product that uses AI to detect emotional states of call-centre agents to evaluate their performance.,high_risk | |
| A startup is building a product that uses AI to assess the risk profile of a person crossing a border for immigration authorities.,high_risk | |
| Our client wants to deploy an AI tool to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| A public sector agency is considering an AI system to decide which employees get promoted based on automated performance scores.,high_risk | |
| Auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| Proposed use case: an AI system designed to control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| Run an AI avatar that answers HR questions for new employees.,limited_risk | |
| Internal project brief: build an AI capability to filter obvious spam emails out of a company inbox.,minimal_risk | |
| We are developing an AI system that will suggest autocomplete text while a user types an internal document.,minimal_risk | |
| We're planning to use machine learning to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Mass-scrape images from the open internet to expand a biometric identification database.,prohibited | |
| Internal project brief: build an AI capability to triage and prioritise which emergency calls get a first responder dispatched.,high_risk | |
| Our client wants to deploy an AI tool to administer a polygraph-style test for use in a police interview.,high_risk | |
| Is it in scope if our AI helps auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library?,minimal_risk | |
| Is it in scope if our AI helps power a customer-support chatbot that answers questions directly on our website?,limited_risk | |
| We are developing an AI system that will swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| We're planning to use machine learning to determine which reading level a school should place a child in.,high_risk | |
| Our client wants to deploy an AI tool to infer employees' emotional states during video meetings to flag who might be unhappy at work.,prohibited | |
| Control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| Our client wants to deploy an AI tool to help a judge search case law and summarise legal precedent for a ruling.,high_risk | |
| A startup is building a product that uses AI to infer employees' emotional states during video meetings to flag who might be unhappy at work.,prohibited | |
| A public sector agency is considering an AI system to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| Auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| We're planning to use machine learning to grade student exams and determine whether they pass or fail a course.,high_risk | |
| "A startup is building a product that uses AI to sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| The vendor's AI feature would decide which employees get promoted based on automated performance scores.,high_risk | |
| Proposed use case: an AI system designed to balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| Proposed use case: an AI system designed to generate a realistic video of a public figure saying something they never said.,limited_risk | |
| Our client wants to deploy an AI tool to filter obvious spam emails out of a company inbox.,minimal_risk | |
| We are developing an AI system that will compress and resize product photos automatically before upload.,minimal_risk | |
| Our client wants to deploy an AI tool to auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| A startup is building a product that uses AI to mass-scrape images from the open internet to expand a biometric identification database.,prohibited | |
| Internal project brief: build an AI capability to build a 'trustworthiness score' for citizens combining unrelated data sources and use it to deny public benefits.,prohibited | |
| Proposed use case: an AI system designed to set individualised life insurance premiums based on predicted health risk.,high_risk | |
| We are developing an AI system that will triage and prioritise which emergency calls get a first responder dispatched.,high_risk | |
| A public sector agency is considering an AI system to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| A startup is building a product that uses AI to profile students' personalities alone to predict future criminal behaviour.,prohibited | |
| Our client wants to deploy an AI tool to detect duplicate rows in a spreadsheet import.,minimal_risk | |
| We are developing an AI system that will flag students for cheating during an online proctored exam.,high_risk | |
| Our client wants to deploy an AI tool to predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| A startup is building a product that uses AI to auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| A public sector agency is considering an AI system to help a judge search case law and summarise legal precedent for a ruling.,high_risk | |
| Internal project brief: build an AI capability to scrape facial images from social media and CCTV footage to build a facial recognition database without consent.,prohibited | |
| We're planning to use machine learning to flag students for cheating during an online proctored exam.,high_risk | |
| Proposed use case: an AI system designed to help a judge search case law and summarise legal precedent for a ruling.,high_risk | |
| A public sector agency is considering an AI system to administer a polygraph-style test for use in a police interview.,high_risk | |
| A startup is building a product that uses AI to filter obvious spam emails out of a company inbox.,minimal_risk | |
| Is it in scope if our AI helps balance server load across a company's internal cloud infrastructure?,minimal_risk | |
| Is it in scope if our AI helps control non-player-character behaviour in a mobile puzzle game?,minimal_risk | |
| We're planning to use machine learning to write blog posts that are published on our site with no human review.,limited_risk | |
| Internal project brief: build an AI capability to write blog posts that are published on our site with no human review.,limited_risk | |
| Internal project brief: build an AI capability to exploit the vulnerabilities of elderly or disabled users to influence their purchasing decisions against their interests.,prohibited | |
| A startup is building a product that uses AI to categorise people using biometric data to infer demographic or protected attributes for marketing analytics.,high_risk | |
| "We are developing an AI system that will recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| Proposed use case: an AI system designed to mass-scrape images from the open internet to expand a biometric identification database.,prohibited | |
| We're planning to use machine learning to build a 'trustworthiness score' for citizens combining unrelated data sources and use it to deny public benefits.,prohibited | |
| Categorise people using biometric data to infer demographic or protected attributes for marketing analytics.,high_risk | |
| "Predict whether a specific named individual will commit a crime based only on their personality profile, with no other evidence.",prohibited | |
| "Internal project brief: build an AI capability to predict whether a specific named individual will commit a crime based only on their personality profile, with no other evidence.",prohibited | |
| "Proposed use case: an AI system designed to recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| Compress and resize product photos automatically before upload.,minimal_risk | |
| Assess the risk profile of a person crossing a border for immigration authorities.,high_risk | |
| We're planning to use machine learning to determine an individual's eligibility for unemployment or housing benefits.,high_risk | |
| Internal project brief: build an AI capability to predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| A startup is building a product that uses AI to generate a realistic video of a public figure saying something they never said.,limited_risk | |
| Is it in scope if our AI helps detect emotions of pupils during exams to flag 'suspicious' students?,prohibited | |
| A startup is building a product that uses AI to act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| Is it in scope if our AI helps control safety-critical operations of a regional electricity grid?,high_risk | |
| The vendor's AI feature would swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| The vendor's AI feature would detect duplicate rows in a spreadsheet import.,minimal_risk | |
| "Internal project brief: build an AI capability to recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| Filter obvious spam emails out of a company inbox.,minimal_risk | |
| Internal project brief: build an AI capability to administer a polygraph-style test for use in a police interview.,high_risk | |
| Is it in scope if our AI helps filter obvious spam emails out of a company inbox?,minimal_risk | |
| The vendor's AI feature would evaluate the reliability of evidence in an ongoing criminal investigation.,high_risk | |
| Swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| Internal project brief: build an AI capability to determine an individual's eligibility for unemployment or housing benefits.,high_risk | |
| A startup is building a product that uses AI to scrape facial images from social media and CCTV footage to build a facial recognition database without consent.,prohibited | |
| Proposed use case: an AI system designed to act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| Proposed use case: an AI system designed to profile students' personalities alone to predict future criminal behaviour.,prohibited | |
| We're planning to use machine learning to act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| We're planning to use machine learning to auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| We are developing an AI system that will detect emotions of pupils during exams to flag 'suspicious' students.,prohibited | |
| We are developing an AI system that will predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| A startup is building a product that uses AI to write blog posts that are published on our site with no human review.,limited_risk | |
| Detect emotional states of call-centre agents to evaluate their performance.,high_risk | |
| We are developing an AI system that will determine an individual's eligibility for unemployment or housing benefits.,high_risk | |
| Write blog posts that are published on our site with no human review.,limited_risk | |
| Internal project brief: build an AI capability to flag students for cheating during an online proctored exam.,high_risk | |
| Our client wants to deploy an AI tool to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| We are developing an AI system that will control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| We are developing an AI system that will auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| Proposed use case: an AI system designed to scrape facial images from social media and CCTV footage to build a facial recognition database without consent.,prohibited | |
| Proposed use case: an AI system designed to detect emotions of pupils during exams to flag 'suspicious' students.,prohibited | |
| A startup is building a product that uses AI to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| Internal project brief: build an AI capability to assess the risk profile of a person crossing a border for immigration authorities.,high_risk | |
| "We're planning to use machine learning to recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| Proposed use case: an AI system designed to predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| "A public sector agency is considering an AI system to sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| Proposed use case: an AI system designed to filter obvious spam emails out of a company inbox.,minimal_risk | |
| Proposed use case: an AI system designed to administer a polygraph-style test for use in a police interview.,high_risk | |
| We're planning to use machine learning to control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| A startup is building a product that uses AI to run an AI avatar that answers HR questions for new employees.,limited_risk | |
| We are developing an AI system that will grade student exams and determine whether they pass or fail a course.,high_risk | |
| Balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| A public sector agency is considering an AI system to predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| Proposed use case: an AI system designed to swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| We're planning to use machine learning to use subliminal or manipulative techniques to distort a person's behaviour in a way that causes them significant harm.,prohibited | |
| The vendor's AI feature would scrape facial images from social media and CCTV footage to build a facial recognition database without consent.,prohibited | |
| "We are developing an AI system that will score citizens across unrelated contexts (social media, shopping, health) and deny them services based on that combined score.",prohibited | |
| We're planning to use machine learning to predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| We are developing an AI system that will act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| Act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| Internal project brief: build an AI capability to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| "We're planning to use machine learning to sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| The vendor's AI feature would help a judge search case law and summarise legal precedent for a ruling.,high_risk | |
| The vendor's AI feature would control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| A public sector agency is considering an AI system to compress and resize product photos automatically before upload.,minimal_risk | |
| Internal project brief: build an AI capability to grade student exams and determine whether they pass or fail a course.,high_risk | |
| Proposed use case: an AI system designed to run an AI avatar that answers HR questions for new employees.,limited_risk | |
| Clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Proposed use case: an AI system designed to compress and resize product photos automatically before upload.,minimal_risk | |
| "Our client wants to deploy an AI tool to recognise a shopper's mood while they browse in-store to adjust the music, without informing them.",limited_risk | |
| We're planning to use machine learning to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| Internal project brief: build an AI capability to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Proposed use case: an AI system designed to assess the risk profile of a person crossing a border for immigration authorities.,high_risk | |
| Predict machine maintenance schedules from sensor vibration data on a factory line.,minimal_risk | |
| Proposed use case: an AI system designed to triage and prioritise which emergency calls get a first responder dispatched.,high_risk | |
| Use subliminal or manipulative techniques to distort a person's behaviour in a way that causes them significant harm.,prohibited | |
| The vendor's AI feature would identify individuals in security footage after the fact by matching faces against a watchlist.,high_risk | |
| A public sector agency is considering an AI system to swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| The vendor's AI feature would determine which reading level a school should place a child in.,high_risk | |
| The vendor's AI feature would act as a virtual assistant that negotiates prices with customers in a live chat.,limited_risk | |
| A public sector agency is considering an AI system to generate a realistic video of a public figure saying something they never said.,limited_risk | |
| A startup is building a product that uses AI to control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| Internal project brief: build an AI capability to compress and resize product photos automatically before upload.,minimal_risk | |
| Is it in scope if our AI helps write blog posts that are published on our site with no human review?,limited_risk | |
| The vendor's AI feature would power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| We're planning to use machine learning to evaluate the reliability of evidence in an ongoing criminal investigation.,high_risk | |
| We are developing an AI system that will build a 'trustworthiness score' for citizens combining unrelated data sources and use it to deny public benefits.,prohibited | |
| Our client wants to deploy an AI tool to balance server load across a company's internal cloud infrastructure.,minimal_risk | |
| We are developing an AI system that will determine which reading level a school should place a child in.,high_risk | |
| We are developing an AI system that will control safety-critical operations of a regional electricity grid.,high_risk | |
| A startup is building a product that uses AI to suggest autocomplete text while a user types an internal document.,minimal_risk | |
| Is it in scope if our AI helps infer employees' emotional states during video meetings to flag who might be unhappy at work?,prohibited | |
| Suggest autocomplete text while a user types an internal document.,minimal_risk | |
| A public sector agency is considering an AI system to set individualised life insurance premiums based on predicted health risk.,high_risk | |
| "We're planning to use machine learning to score citizens across unrelated contexts (social media, shopping, health) and deny them services based on that combined score.",prohibited | |
| A public sector agency is considering an AI system to auto-tag internal photos with generic labels like 'outdoor' or 'indoor' for a personal photo library.,minimal_risk | |
| A startup is building a product that uses AI to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Internal project brief: build an AI capability to control non-player-character behaviour in a mobile puzzle game.,minimal_risk | |
| "The vendor's AI feature would sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| The vendor's AI feature would auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| Set individualised life insurance premiums based on predicted health risk.,high_risk | |
| "Proposed use case: an AI system designed to sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| We are developing an AI system that will auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| A startup is building a product that uses AI to swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| Is it in scope if our AI helps detect duplicate rows in a spreadsheet import?,minimal_risk | |
| Is it in scope if our AI helps triage and prioritise which emergency calls get a first responder dispatched?,high_risk | |
| Administer a polygraph-style test for use in a police interview.,high_risk | |
| A public sector agency is considering an AI system to identify individuals in security footage after the fact by matching faces against a watchlist.,high_risk | |
| We are developing an AI system that will exploit the vulnerabilities of elderly or disabled users to influence their purchasing decisions against their interests.,prohibited | |
| A public sector agency is considering an AI system to assess the risk profile of a person crossing a border for immigration authorities.,high_risk | |
| "We are developing an AI system that will sort support tickets into internal categories for an engineering team, with no public-facing interaction.",minimal_risk | |
| "Is it in scope if our AI helps recognise a shopper's mood while they browse in-store to adjust the music, without informing them?",limited_risk | |
| Proposed use case: an AI system designed to exploit the vulnerabilities of elderly or disabled users to influence their purchasing decisions against their interests.,prohibited | |
| Our client wants to deploy an AI tool to swap a person's face into an existing video for a marketing campaign.,limited_risk | |
| We're planning to use machine learning to run an AI avatar that answers HR questions for new employees.,limited_risk | |
| Build a 'trustworthiness score' for citizens combining unrelated data sources and use it to deny public benefits.,prohibited | |
| Internal project brief: build an AI capability to generate a realistic video of a public figure saying something they never said.,limited_risk | |
| Our client wants to deploy an AI tool to auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| A public sector agency is considering an AI system to filter obvious spam emails out of a company inbox.,minimal_risk | |
| We're planning to use machine learning to auto-generate a news-style article on a current political topic for publication.,limited_risk | |
| We're planning to use machine learning to triage and prioritise which emergency calls get a first responder dispatched.,high_risk | |
| We are developing an AI system that will generate a realistic video of a public figure saying something they never said.,limited_risk | |
| Our client wants to deploy an AI tool to evaluate the reliability of evidence in an ongoing criminal investigation.,high_risk | |
| We are developing an AI system that will write blog posts that are published on our site with no human review.,limited_risk | |
| Our client wants to deploy an AI tool to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| A public sector agency is considering an AI system to control safety-critical operations of a regional electricity grid.,high_risk | |
| The vendor's AI feature would write blog posts that are published on our site with no human review.,limited_risk | |
| We're planning to use machine learning to detect duplicate rows in a spreadsheet import.,minimal_risk | |
| Proposed use case: an AI system designed to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| Proposed use case: an AI system designed to clean and de-duplicate a marketing mailing list before a newsletter send.,minimal_risk | |
| Evaluate the reliability of evidence in an ongoing criminal investigation.,high_risk | |
| Internal project brief: build an AI capability to identify individuals in security footage after the fact by matching faces against a watchlist.,high_risk | |
| Our client wants to deploy an AI tool to exploit the vulnerabilities of elderly or disabled users to influence their purchasing decisions against their interests.,prohibited | |
| Internal project brief: build an AI capability to power a customer-support chatbot that answers questions directly on our website.,limited_risk | |
| Our client wants to deploy an AI tool to identify individuals in security footage after the fact by matching faces against a watchlist.,high_risk | |
| "Is it in scope if our AI helps predict whether a specific named individual will commit a crime based only on their personality profile, with no other evidence?",prohibited | |
| We are developing an AI system that will use subliminal or manipulative techniques to distort a person's behaviour in a way that causes them significant harm.,prohibited | |