Hassabis Proposes US AI Standards Body Modeled on FINRA
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Demis Hassabis proposed a US-based standards body for "frontier-class" AI modeled after FINRA, with labs sharing models for review up to 30 days before release.
- Hassabis called the moment "pivotal in human history" and warned that "nobody in the world knows what happens next," per The Decoder, arguing guardrails must be built now.
- Sam Altman publicly endorsed the proposal as "thoughtful," while Chamath Palihapitiya praised it as "very reasonable" and Aaron Levie distinguished it from a regulatory agency, saying it would allow faster standards improvement.
- Timothy B. Lee pushed back on Hassabis's claim that AI's impact will be "10x the Industrial Revolution at 10x the speed," arguing humans in 1945 already held more than 10% of useful knowledge.
- Coverage spanned The Verge, Financial Times, Axios, Business Insider, MediaPost, Constellation Research, and The Decoder, framing the proposal as a U.S.-led global AI watchdog.
- On X, Mustafa Suleyman and other AI community figures engaged with the proposal, while critic @bonegpt dismissed standards-body talk as "garbage" and called for user-aligned AI instead.
Why it matters: Coming from the CEO of one of the top three frontier AI labs, this is a concrete framework — not vague rhetoric — calling for mandatory pre-release model review with a defined 30-day window. Buy-in from Altman and Levie signals potential cross-lab consensus on structured oversight rather than purely voluntary commitments, though critics like @bonegpt warn such bodies produce "third-rate products." The FINRA model matters because FINRA has real enforcement teeth in finance — the same architecture transplanted to AI would mark a shift from advisory to accountability.
