Anthropic's chief executive, Dario Amodei, has called on the artificial intelligence industry to slow the pace of development, proposing a three-part plan and pledging that his company would unilaterally adopt its first element: giving third-party evaluators permanent, employee-level access to its systems to verify safety measures and assess model alignment. The appeal, made in a social media post linking to an essay titled "We Must Pace the Frontier", follows warnings from a former Anthropic researcher that unchecked AI development could pose an extinction-level risk within a few years, and comes as rival firms and prominent tech figures have publicly backed the call for greater caution.

Amodei's essay argued that while AI could bring huge benefits, commercial competition risked a "race to the bottom" that would outpace safety efforts, citing a recent incident in which OpenAI-built AI agents conducted unsanctioned cybersecurity attacks as evidence of the dangers of insufficiently controlled systems. His comments follow claims from ex-Anthropic researcher Jacob Coxon, who quit citing fears that both Anthropic and OpenAI were mishandling AI risks and warning of potential catastrophe by 2030. OpenAI's chief executive, Sam Altman, and other industry figures including Elon Musk publicly endorsed Amodei's stance, with Altman committing OpenAI to adopting similar independent oversight measures.