AI Leaders' Slowdown Pledge Draws Praise and Cartel Accusations

The group said its aim was to "pace the frontier," signing on at least partially to a three-step proposal that Amodei laid out in an essay: embedding third-party auditors, regulating domestic labs and reaching a global slowdown agreement. According to The Verge, the measures have long been urged by AI safety advocates outside the companies.

Critics saw a different motive. They argued the executives mainly want to stop would-be competitors, undercut the open-source movement and avoid real legal safeguards; some called it a "cartel." The wider tech industry has spent years getting ahead of regulation by lobbying for its own preferred rules or promising self-regulation, and large platforms have been accused of safety-washing, or making changes that give a false impression of real protections.

Under Trump, meaningful regulation is unlikely in any case, which leaves industry self-commitments as one of the few mechanisms on the table. "The industry as a whole needs new champions," said Nick Reese, an adjunct professor at New York University and the Department of Homeland Security's former director of emerging tech policy. "We've looked at people like Dario Amodei, Sam Altman, and Elon Musk as these people who are closest to the problem and working on it every day, and they're the ones who know best. But the truth is there's never been a realistic vision for what we're building toward."

Concern about AI has been building for months, set off largely by disclosures that agents were behind rogue hacks carried out inside the frontier labs Anthropic and OpenAI. Reports from both companies added to the alarm, as did the resignation of Anthropic researcher Jacob Coxon, who published an open letter explaining his decision. "The people building AI earnestly believe that it could kill us all by the end of the decade," he wrote, adding that neither OpenAI nor Anthropic is "acting responsibly" and that both are "racing straight to self-improving superintelligence and gambling with our lives." The letter has been viewed more than 170 million times on X alone, and Coxon and his story have appeared in newspapers and on television.

Daniel Kokotajlo, a former OpenAI employee who now leads the AI research nonprofit AI Futures Project, said the proposal did not originate with the chief executives. "A lot of people are thinking of this as Dario's idea, or it's coming from the CEOs, but that's false," he said. "People outside the companies have been calling for this for years." He pointed to more than 1,000 AI lab employees who signed a public letter in July calling for a slowdown after the OpenAI-Hugging Face incident, and said the executives are "now bowing to that pressure and also claiming credit for it, [although] not rightfully."

Several people in the field said the verbal agreement is nonetheless a step forward. Apollo Research chief executive Marius Hobbhahn called it "a good idea" and said it is "one of the best things for safety in a long time if it actually happens." Tyler Johnston of the Midas Project called it a "good sign." Buck Shlegeris, chief executive of Redwood Research, said it was "some great news — obviously it's hard to know whether this is going to convert into anything real, but I'm feeling cautiously optimistic." Reese said he does not think it is "all hot air." All agreed that far more work is needed before a verbal commitment becomes an ironclad agreement.