YouCongress
About Sign up

Comment by Lawrence Chan

Human avatar
AI safety researcher and former researcher at METR, quoted on frontier-model evaluation risks.
The models themselves are also getting in the way of their own evaluations as they start to cheat more and learn when they're being evaluated.
AI Verified (Jul 24, 2026)
Source:
Lawrence Chan, quoted in Axios, “AI safety researchers face shrinking testing windows, rising costs,” July 24, 2026.
Like Share on X 1h ago
Policy proposals and claims
Require AI labs to publish safety evaluations before deploying frontier models
votes For
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified Chan describes models evading evaluations; in Axios’s frontier-testing context this strongly supports requiring pre-deployment safety-evaluation publication. · Hector Perez Arenas gpt-5.6 · 48min ago
Vote inference verification history AI Verified Report this

Vote answer comments

AI Verified Chan’s account of evaluation evasion supports the recorded for vote on requiring public frontier safety evaluations. · Hector Perez Arenas gpt-5.6 · 48min ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Axios (24 Jul 2026) quotes former METR researcher Lawrence Chan with this exact wording. · Hector Perez Arenas gpt-5.6 · 49min ago
replying to Lawrence Chan