Bill Gurley Criticizes Anthropic CEO's AI Regulation Proposal, Citing Regulatory Capture

On September 16, prominent tech investor Bill Gurley criticized Anthropic CEO Dario Amodei's proposal for AI regulation on social media platform X, arguing that regulation is difficult to design correctly and highly vulnerable to "regulatory capture."

Amodei had published a long post on X on September 12, calling for slowing the pace of frontier AI model improvements to give safety measures more time to catch up. He also suggested that independent third-party assessment teams should be granted employee-level access to review AI companies' safety practices, report incidents, and evaluate models and training processes. His stance was supported by OpenAI CEO Sam Altman and Elon Musk.

Gurley said he dislikes regulation because it is hard to design properly and is easily captured by regulated industries. He pointed to the two Boeing 737 MAX crashes and the Fukushima nuclear accident as cases where regulatory capture was a core failure, citing the relationships between Boeing and the Federal Aviation Administration (FAA), and between Tokyo Electric Power Company (Tepco) and Japan's Nuclear and Industrial Safety Agency (NISA). He argued that if safety is the top priority, regulators must be in a position of power and accountability above the companies they oversee, and that any structure allowing close ties between regulators and regulated firms is fundamentally flawed.

Gurley also referenced Kiyoshi Kurokawa, who led Japan's parliamentary investigation into the Fukushima disaster and viewed regulatory capture as a key institutional cause.

The renewed public debate over AI safety has intensified discussions on whether AI development should be constrained to control risks, while also heightening concerns about potential existential threats from AI. Okta co-founder and CEO Todd McKinnon said pessimistic predictions about AI are extreme, adding that AI brings both risks and opportunities, and that a balance must be found.