license: cc-by-4.0
language:
- en
pretty_name: 'The Machine Record: incidents involving commercial physical robots'
size_categories:
- n<1K
task_categories:
- tabular-classification
- text-classification
tags:
- robotics
- robot-safety
- autonomous-vehicles
- robotaxi
- humanoid-robots
- incidents
- recalls
configs:
- config_name: default
data_files:
- split: train
path: data/cases.csv
- config_name: events
data_files:
- split: train
path: data/events.csv
- config_name: sources
data_files:
- split: train
path: data/sources.csv
- config_name: sgo_links
data_files:
- split: train
path: data/sgo_links.csv
The Machine Record: incidents involving commercial physical robots
A mirror of the public register at themachinerecord.com. Each row is a documented incident involving a commercial physical robot (robotaxis, delivery robots, warehouse and industrial machines, surgical systems, humanoids), with a permanent case page that carries the timeline from first report through any recall, investigation, lawsuit or settlement, and the archived sources.
- Four tables: cases (24 columns), timeline events, sources with tiers and archived copies, links to NHTSA Standing General Order filings. Codebook, annotation guidelines and limitations on GitHub: https://github.com/vzolot/machine-record-data
- Refreshed daily from the site's open API
- Not a safety rating: fleet sizes are not public, so rates cannot be computed honestly from this file
Licence
Case metadata is CC BY 4.0, attribution to The Machine Record (themachinerecord.com). Quoted text from news sources is not included and stays with its publisher.
Citation
DOI for the dataset (all versions): https://doi.org/10.5281/zenodo.22925531
Cite a case by its reference and URL (TMR-0238, https://themachinerecord.com/case/...),
or the dataset as a whole by the DOI above; CITATION.cff in the GitHub repository carries the same reference.