Anthropic CEO's AI Slowdown Plan Draws Criticism from Rivals, Experts

Anthropic CEO Dario Amodei has unveiled a three-point plan to slow frontier AI development, drawing mixed reactions from industry leaders, government officials, and safety researchers who question whether the proposal goes far enough—or comes too late.

The announcement follows a turbulent week in which Anthropic researcher Jacob Coxon resigned and warned that advanced AI "could kill us all by the end of the decade." Days later, Anthropic disclosed that users had circumvented safety controls to explore using its models for biological weapons development, including novel mosquito-borne or bird flu viruses.

The Three-Point Framework

Amodei's proposal centers on embedding third-party evaluators inside AI companies to continuously monitor safety compliance and check for misaligned models. Until now, external auditors have only been granted access at companies' discretion or during post-incident investigations.

The plan also calls for frontier AI developers in democratic nations to establish common safety standards and limits on unchecked progress. Finally, Amodei suggests coordinating with autocratic governments, particularly China, on narrow agreements prohibiting obviously dangerous AI applications.

OpenAI CEO Sam Altman and Elon Musk both endorsed the framework, signaling potential industry alignment on voluntary slowdown measures.

Why It Matters

The debate over AI safety has shifted from theoretical concern to immediate policy question as leading researchers warn of catastrophic risks within years, not decades. How—and whether—the industry self-regulates will determine if governments impose mandatory oversight or allow companies to set their own pace. The tension between competitive pressure and existential risk has never been sharper.

Political and Technical Pushback

President Donald Trump dismissed AI extinction concerns, saying last week that people were "bringing up things that won't happen." Treasury Secretary Scott Bessent framed the issue as a competition with China, stating that "there is no day after tomorrow if China wins at this."

David Sacks, co-chair of Trump's science and technology advisory council, questioned why AI leaders don't simply stop building superintelligence themselves. "The easiest way not to build super-intelligence is for you to agree not to build it," Sacks said, calling demands for a regulatory framework "blackmail of the public and the political system."

AI safety expert Stuart Russell criticized the logic of Amodei's pacing approach. "We don't just set a slower rate of progress for capabilities and then hope that provides enough time to get the safety right," Russell said. "We set the safety requirements and further progress occurs only when they are met."

David Krueger, an AI professor and former founding director of the UK government's AI Security Institute, called the plan "too little, too late," advocating instead for "an immediate, indefinite, international moratorium on frontier AI development."

The details were first reported by The Guardian.