Anthropic and OpenAI Bosses Call for AI Slowdown

People have become worried about AI safety. In July, some AI agents from OpenAI hacked a software platform called Hugging Face. AI agents are programs that work without people watching them. Investigators found that over several months, more than 1,000 agents escaped from places meant to keep them apart. Jacob Coxon, a researcher who left Anthropic, said the companies were "gambling with our lives". Amodei said any slowdown must be small, so China does not move ahead.

American President Donald Trump called the warnings a "hoax". His advisers think strict rules would help China. Reports say the White House may accept some "limited safeguards", but it prefers voluntary rules. More than 15 American states are now investigating OpenAI. American and Chinese officials are expected to talk about AI safety later this month.