Google, OpenAI, Anthropic coordinate on "Standards Authority for Frontier AI"

The move by Google, OpenAI, and Anthropic to coordinate on a "Standards Authority for Frontier AI" comes as the two latter firms are probing "tens of thousands" of security incidents linked to their models. The context is a wider gap: frontier AI development is outpacing governments' capabilities and willingness to police it, and a former EU official pointed to a "disconnect between the law and the reality of what is being regulated." The EU's 2024 AI Act, once touted as the first comprehensive AI legislation, has been overtaken by technological developments, The New York Times wrote, calling it a "cautionary tale." In the US, President Donald Trump has opposed regulation, though Transformer wrote it is "probably going to happen anyway." Against that backdrop, the three labs' self-organized standards body could be read as an attempt to shape rules before regulators write them. Whether it carries weight beyond its members is likely to hinge on how transparently OpenAI and Anthropic share the results of their incident probe and whether other AI developers sign on.