Regulatory capture and the case for independent AI oversight

The post’s author is skeptical of regulation, calling it difficult to get right and vulnerable to regulatory capture. The author says investigations into the 737 Max crashes and Fukushima accident identified capture as the failure behind those incidents. Applied to AI safety, the argument is that oversight cannot rest with a friendly third party with close industry ties. Regulators should instead be fully independent, devise their own tests and procedures, and have authority over companies while remaining accountable to the American public.