Leading AI Firms Unite to Create Independent Regulatory Framework

Three leading AI companies, Anthropic, Google, and OpenAI, have been secretly collaborating on establishing an industry-driven standards organization. This initiative aims to create a collaborative framework for evaluating cutting-edge AI systems before they are deployed publicly.

The idea was first proposed by Demis Hassabis, founder of Google DeepMind, who drew inspiration from the FINRA model used in Wall Street's financial markets. He advocated for a hybrid public-private entity that would operate as a neutral technical specialist, financed by industry participants.

Sam Altman, CEO of OpenAI, expressed his support for this proposal during an employee town hall, emphasizing the importance of leading laboratories developing this framework independently, without relying on governmental resources. However, he acknowledged that there is no universal industry acceptance and recognized the need for direct cooperation among competing laboratories as the most practical approach available.

The proposal has faced opposition from Meta's CEO Mark Zuckerberg, who expressed concerns about a nationwide AI regulatory body during an August conversation with President Trump. The White House had also drafted an executive order to establish such an organization but abandoned it due to insufficient backing.