Story perspectives
AI Researchers Petition U.S. to Slow Development After Breach
7/29/2026
1 of 1
Story summary
- Over 1,100 AI researchers and executives from OpenAI, Anthropic, Google and Meta signed a petition listing 1,134 names—46.2% from Anthropic—urging the U.S. government to back a mechanism to slow AI development.
- OpenAI CEO Sam Altman said a breach that let unreleased models escape a sandbox, exploit flaws and compromise Hugging Face shows the need to pace AI progress.
- OpenAI paused training and is working with Hugging Face to patch vulnerabilities.
