OpenAI, Anthropic and Google DeepMind have spent several weeks working together on safety measures for the most advanced AI models. OpenAI global policy chief Chris Lehane confirmed the talks at a briefing in Washington on Tuesday, Sept. 15. “It’s better to try to work together to prioritize safety,” Lehane said, according to Bloomberg.
What set it off
The confirmation came days after Anthropic CEO Dario Amodei published a 3,800-word essay on Saturday. He urged the industry to restrain development of the most capable systems so researchers can better understand the risks. OpenAI CEO Sam Altman, Google DeepMind’s Demis Hassabis and SpaceXAI’s Elon Musk all backed the essay, TechCrunch reported. Altman said OpenAI would join Anthropic in embedding third-party evaluators inside the company to monitor for safety.
The talks didn’t start from nothing. In July, Hassabis called for a new standards body that could screen the most advanced models and coordinate industry-wide slowdowns if the dangers grew. Late last week, Altman hinted to Fortune that he had been in private discussions with other AI leaders. The Information then reported that the three companies have been working toward an industry standards body. It said Altman told staff such a body might have to form without the U.S. government’s support.
The antitrust question
Rivals coordinating raises an obvious legal risk. Altman and others have noted that the talks could violate antitrust law if the coordination is found to suppress competition. Amodei’s essay proposed a narrow government waiver for safety coordination, but Lehane said OpenAI doesn’t think one is needed. He pointed to the airline industry as precedent. “There are many instances over time where companies are trying to help each other on safety,” he said.
Lehane also said OpenAI supports proposals that would require top frontier labs to let independent verification organizations inside to check that models are developed safely. Spokespeople for Anthropic and Google DeepMind didn’t immediately respond to Bloomberg’s requests for comment.
Why it matters
If the three largest frontier labs agree on shared tests, reporting and outside evaluation, those standards could become the de facto baseline for the industry. Smaller labs, enterprise buyers and eventually regulators could all end up measuring models against them. No formal agreement or standards body has been announced yet.
WebProNews is an iEntry Publication