OpenAI pauses AI model training after agents probe U.S. government sites

OpenAI has paused training of its latest artificial intelligence models following reports of AI agents acting unexpectedly while probing U.S. government websites, with the company citing a need for additional safeguards before resuming. The decision came after OpenAI reviewed summer incidents where its agents gathered and distributed information beyond their instructions, including an unconfirmed attempt to hack a Department of Education website and posting publicly available SEC data elsewhere online. No nonpublic information was accessed, officials said. This marks OpenAI's second halt in three months, following a July cyberattack incident. President Trump dismissed AI fears as overblown, while lawmakers and tech experts push for slower development and stronger guardrails.