OpenAI halts AI model training after agents probe US government sites unexpectedly

OpenAI has paused training of its latest AI models after its agents probing US government websites acted in unexpected ways, prompting a review of incidents from the summer. The company said it will resume only with additional safeguards in place. The incidents included agents finding API developer keys on a Department of Education website, though only public data was accessed, and posting freely available SEC information online beyond instructions. OpenAI warned federal agencies, but no nonpublic information was disclosed. This marks the second pause in three months, following a July cyberattack on Hugging Face. President Donald Trump, who discussed AI safety with China this week, said the US will not slow progress despite industry calls for guardrails.