Demis Hassabis proposes AI standards body modeled on FINRA

Google DeepMind CEO Demis Hassabis has proposed the creation of a U.S. Frontier AI Standards Body, modeled after the Financial Industry Regulatory Authority (FINRA), to review advanced AI models before their release. Hassabis stated that frontier labs would voluntarily share models with the standards body for review up to 30 days before release, with the potential for mandatory pre-release assessments for the most capable systems.

Hassabis warned that artificial general intelligence (AGI) is likely only a few years away and could reshape human civilization, comparing its impact to the discovery of electricity or fire. He expressed concern that AI capabilities are advancing faster than society's ability to manage the risks, which could include cybersecurity, biological, nuclear, and national security threats.

The proposal follows similar calls for AI oversight from industry leaders, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei. However, White House AI advisor Sriram Krishnan has previously dismissed the idea of an 'FDA for AI.' Hassabis emphasized the need for robust safeguards and a societal effort to shape AGI for the benefit of humanity.