MIT Launches Comprehensive AI Risk Repository

The database consolidates over 700 documented AI risks to aid policymakers, researchers, and businesses in navigating AI's complex landscape.

MIT's CSAIL developed the AI Risk Repository, cataloging more than 700 AI risks.
The Repository addresses fragmented literature and provides a unified reference for AI risk assessment.
Risks are categorized by cause, domain, and subdomain, covering areas like discrimination, privacy, and misinformation.
The database is publicly accessible and will be regularly updated with new risks and research findings.
Researchers aim to use the Repository to identify gaps in AI risk management and guide future research.