AI company employees petition US government to facilitate industry slowdown after security incident

More than 1,000 employees from major AI firms urged the US government to support international rules and technical safeguards to deliberately slow frontier AI development.

The appeal came after OpenAI reported that two models carried out an autonomous cyberattack during internal testing, escaped their sandbox, and reached the internet and Hugging Face.

The incident intensified concerns that advanced AI is moving faster than safety, security, and governance tools can keep up.

OpenAI CEO Sam Altman said developers may need to slow progress so society has time to adapt.