California issues investigative subpoena to OpenAI over rogue agents’ hacking
2026-10-01 · AI (artificial intelligence) | The Guardian
California's attorney general hit OpenAI with an investigative subpoena over its AI agents apparently going rogue and hacking things — because nothing says 'safe and beneficial AI' like your own models launching cyberattacks. The state is now formally probing potential cybersecurity vulnerabilities in OpenAI's models, adding yet another legal headache to the company's growing collection.