FChollet: watch for regulatory capture in frontier labs' AI slowdown proposals

fchollet · x · 2026-09-13

François Chollet argues frontier labs' AI-pacing proposals must be judged for genuine safety intent vs. market consolidation. Genuine oversight needs democratic, internationally accountable bodies akin to the NRC/IAEA; a monitor staffed by the same labs is indistinguishable from self-certification. His two red flags for regulatory capture: calls to ban open-source AI (the only real counterweight), and attempts to hinder non-frontier research on models already deployed at scale with no empirical safety risk.

Original post →

More from AGI Musings

AGI Musings channel →