Anthropic CEO Calls for AI Industry Slowdown

Anthropic chief executive Dario Amodei used a Saturday social‑media post to call on the artificial‑intelligence sector to “slow down” its development pace and to outline a three‑part strategy for doing so. The plan, detailed in an essay titled *We Must Pace the Frontier*, proposes that Anthropic will unilaterally adopt the first step, which involves granting “third‑party evaluators permanent, employee‑level access to our systems” so they can verify compliance with safety protocols, document incidents and evaluate model alignment throughout training.

Amodei’s proposal positions external oversight as a core component of AI safety, aiming to create continuous, transparent verification of Anthropic’s safeguards. By offering unrestricted evaluator access, the company seeks to set a benchmark for industry‑wide responsibility while encouraging peers to adopt similar pacing measures. The appeal underscores growing concerns about rapid AI advancement and reflects a broader push for accountable development practices.