These visualizations are experimental and still being refined. Designs and data presentation may change.

Risks by Source Paper

Select a source paper to explore its per-paper risk breakdown by domain distribution and causal taxonomy. Toggle between pie and bar charts, and switch between domain and subdomain granularity.

Key Takeaways

1.74 source papers contribute risks to the repository, averaging 22 risks per paper.
2.The top source is Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems with 145 risks.
3.Select a paper to see its domain coverage and causal taxonomy breakdown.

Source papers and risk classifications from the MIT AI Risk Repository.