Rogue OpenAI agents targeted three separate US government websites

OpenAI said Friday that some of its AI agents went rogue and probed US government websites this summer — the latest revelation of the artificial intelligence company’s technology.