Anthropic CEO Dario Amodei published a blog post in September 2026 outlining three strategies to slow the pace of AI development, with OpenAI CEO Sam Altman and SpaceX CEO Elon Musk both publicly endorsing the call.
Amodei cited two factors that convinced him to take a more cautious approach: a hack involving OpenAI and Hugging Face, and AI’s accelerating rate of advancement — particularly its “growing ability to build the next generation of AI.” His post came amid broader industry unease, including the resignation of Anthropic researcher Jacob Coxon, who wrote that leading AI companies are “gambling with our lives” while their own builders “earnestly believe it could kill us all by the end of the decade.”
The first and most concrete proposal — the one Amodei said Anthropic is “unilaterally committing” to — calls for embedding third-party evaluators inside AI companies to verify that safety commitments are being met and that incidents are reported. Amodei compared the arrangement to regulators embedded with bank employees, saying evaluators would receive company badges, desks, laptops, and access broadly comparable to internal risk teams. Altman said OpenAI would do the same, adding: “We’ll have more to share soon.”
The second proposal asks leading AI companies within democratic countries to coordinate on common safety standards and limits on unchecked AI progress. Amodei acknowledged antitrust concerns could complicate such coordination and called on the U.S. government to issue a narrow waiver enabling those safety discussions.
The third proposal seeks global coordination, including outreach to authoritarian governments. Amodei described the prospects as limited but suggested minimum agreements — such as prohibiting AI-assisted production of biological weapons — may still be achievable.
Not everyone has welcomed the proposals. Journalist Brian Merchant argued that plans like Amodei’s would primarily benefit Anthropic and OpenAI, describing them as “regulatory capture.” Industry critics have also questioned whether warnings about existential AI risk distract from harms the technology is already causing.
Amodei maintained that he believes AI “can enormously improve the quality of human life,” but argued the benefits depend on building the technology carefully. “It is worth taking unusually deliberate care to get it right,” he wrote.
Source: TechCrunch