Google, Anthropic, and OpenAI Explore Creating a Body to Set Common AI Standards

Anthropic, OpenAI, and Google, a subsidiary of Alphabet, are in discussions to establish an industry-led body to set common standards for testing and evaluating artificial intelligence, a move that could reshape how leading models are assessed before regulators impose stricter rules. These talks carry significant weight for investors, as shared standards could increase development costs and slow deployment, yet they may also reduce regulatory uncertainty for companies investing heavily in AI. According to The Information, the discussions are ongoing, with meetings held last week. The conversations began before Anthropic CEO Dario Amodei called on AI companies last Saturday to coordinate on testing and evaluation and to slow the pace of AI development. OpenAI CEO Sam Altman also told employees that he supports creating such an organization, although he reportedly believes major AI labs will need to establish it without support from the U.S. government. This proposal would effectively give leading AI developers a greater role in determining how increasingly powerful models are tested for safety, reliability, and other risks.