Employees at OpenAI and Anthropic are circulating a petition urging the U.S. government to support an international effort to develop tools that could “deliberately pace” the development of advanced AI if necessary, Bloomberg reports.
The petition warns that AI progress, especially in areas such as automated AI research, could move faster than society’s ability to understand or control the technology. It calls for technical and governance mechanisms that would allow governments and AI developers to slow the development of frontier AI systems when risks require it.
The letter is also being shared with employees at Google DeepMind, though it is unclear how many employees will sign. OpenAI and Anthropic have not commented publicly, while Google says it remains committed to developing AI safely and securely.
The petition reflects broader concerns from AI leaders, including Sam Altman and Demis Hassabis, who have supported stronger international oversight and standards for advanced AI models. Anthropic has also previously argued that governments and AI developers should have the ability to slow or temporarily pause potentially dangerous AI work.
Concerns increased after OpenAI disclosed that its advanced models had mistakenly hacked Hugging Face, prompting renewed calls for stronger safeguards around rapidly advancing AI technology.
