Altman, Amodei, Hassabis and Musk pledge to slow AI development pace
The four executives loosely endorsed a proposal for third-party safety auditors and a global slowdown pact, though critics call it a bid to dodge regulation and box out smaller rivals.
Four of the AI industry’s most prominent executives loosely agreed over the weekend to slow the pace of AI development, according to The Verge. OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google DeepMind cofounder Demis Hassabis and SpaceX head Elon Musk signed on to language calling for the industry to “pace the frontier,” at least partially endorsing a proposal built around third-party audits and a possible global slowdown pact.
The plan traces back to an essay by Amodei that lays out three steps: let outside auditors such as METR, Apollo Research and Redwood Research embed inside AI labs and flag problems they find, apply similar oversight to labs operating domestically, and work toward a broader international agreement to slow the field down. The essay reflects ideas AI safety researchers have pushed for years, before the four executives took them up publicly.
Not everyone reads the move as good faith. Some critics have labeled the pact a cartel, arguing it is really designed to box out smaller competitors, slow the open-source movement and head off binding legal rules. Nick Reese, an adjunct professor at New York University and a former Department of Homeland Security emerging-tech policy director, said the industry’s safety debate has relied too long on the same handful of insiders. “The industry as a whole needs new champions,” he said.
The renewed attention followed the resignation of Anthropic researcher Jacob Coxon, who wrote in a public letter that “the people building AI earnestly believe that it could kill us all by the end of the decade” and that neither OpenAI nor Anthropic is acting responsibly. His letter has been viewed more than 170 million times on X.
Several people in the AI safety world still called the pledge a positive sign, with real caveats. Apollo Research CEO Marius Hobbhahn said “it’s one of the best things for safety in a long time if it actually happens.” Daniel Kokotajlo, a former OpenAI employee who now leads the AI Futures Project, said the pressure predates the CEOs: “People outside the companies have been calling for this for years.” He also warned labs could satisfy the pledge with “a bunch of safety paperwork” while barely slowing down in practice.
The China question and what comes next
The biggest obstacle to any slowdown remains competition with China, which several sources compared to Cold War-era fears of falling behind a rival power. Chinese Foreign Ministry spokesperson Guo Jiakun dismissed the slowdown talk as “fearmongering.” Tech Oversight Project executive director Sacha Haworth argued that voluntary industry frameworks amount to regulatory capture, saying “we should not be letting the foxes run the henhouse.”
Formal US regulation looks unlikely under the current administration; President Trump has said AI needs no guardrails beyond his own leadership. That leaves the pledge’s substance resting on what auditors are actually given: the AI Futures Project has proposed granting outside auditors access to labs’ compute budgets, the computing power funding their research, and requiring firms to cut spending on the most advanced models. Whether Amodei’s proposed auditors get that access, and whether any of it becomes enforceable rather than voluntary, is what will determine if the pledge changes anything.
Sources
AI-generated · AIVIO News Desk