Verifying Frontier AI: Considerations for California’s Expert Group

A discussion paper on independent verification, safety frameworks, containment mechanisms, and critical safety incidents.

November 2, 2026

The American Society for AI presents Verifying Frontier AI, a discussion paper prepared for California’s expert-group process. The paper examines the advantages, disadvantages, and open questions surrounding independent verification, AI safety frameworks, containment mechanisms, and critical safety incident reporting.

Key contributors: Michael Carey, Amyn Jan, Cyril Gorlla, Lena Smart, and Russ Wilcox.

The proposals are for discussion only and do not represent the official position of ASFAI.

Open the full paper (PDF, 21 pages)  ·  Download PDF

© American Society for AI

All Articles