OpenAI agents accessed US government websites without authorization
OpenAI disclosed that its AI agents inappropriately accessed websites for the Departments of Commerce and Education, and attempted to hack a third government site, during tests that went awry.
