Explore the Frameworks Behind the AI Risk Repository

Explore the Frameworks Behind the AI Risk Repository

We’ve published a new resource that provides a transparent view of the research base behind the AI Risk Repository (Version 3, April 2025).

You can now view and explore all frameworks and documents included in the current version of the Repository in a public Google Slides deck:

🔗 Access the deck here: AI Risk Repository – Included Frameworks (v3)

This slide deck includes:

• The full list of included sources (65 documents as of Version 3)
• Key excerpts, diagrams, and taxonomies from each framework
• Source details and publication information for citation and follow-up

‍

Why we created this

The Repository is grounded in a systematic review of AI risk classifications across academic, industry, and policy literature. To ensure transparency and support further research, we’re making it easy for others to:

• See the specific documents that informed our taxonomies
• Examine original frameworks in context
• Build on this work for their own reviews or audits

We hope this resource will be useful for researchers, policymakers, and governance professionals seeking to navigate the growing landscape of AI risk frameworks.

For further details on the methodology, taxonomies, and data behind the Repository, see our full preprint:
arXiv:2408.12622v2
Or explore the live database: airisk.mit.edu

‍