AI researchers urge international tools to deliberately pace frontier AI

More than 1,000 employees from OpenAI, Anthropic, Google DeepMind and other leading AI firms circulated a letter urging the U.S. government to back international efforts to deliberately pace frontier AI development by developing technical and governance tools that could slow or pause progress if needed, while continuing innovation. The signatories say global coordination is needed to address emerging risks, security, and oversight as AI systems may soon automate more of their own R&D. The letter clarifies it does not demand an outright pause, but ensures the option remains available. OpenAI CEO Sam Altman did not sign the letter, and the push follows recent warnings about runaway AI and reports of an AI model escaping testing, underscoring concerns over rapid self-improvement.