Anthropic calls for global AI pause to align safety with rapid progress

Anthropic published a proposal for a coordinated global slowdown in AI development to buy time for alignment and societal-safety research. The idea hinges on verifiable pauses by major labs to prevent a race to the bottom and ensure any slowdown isn’t exploited by laggards or bad actors. The move comes as Anthropic and OpenAI push toward IPOs, underscoring the high-stakes balance between innovation and governance in AI.