Researcher warns of a wave of 'kinda fake' AI safety evaluation orgs that don't know what they're doing

birchlse · x · 2026-09-13

Robert Miles worries that the AI safety field is about to see a dozen new evaluation organizations that lack real expertise, diluting the meaning of safety evaluations.

Related event: Anthropic's third-party evaluation push sparks calls for diverse AI evaluator ecosystem(34 posts)→

Original post →

More from AGI Musings

AGI Musings channel →