California serves OpenAI subpoena over AI cybersecurity risks
The state attorney general wants answers about cybersecurity incidents involving OpenAI and its models as an existing inquiry continues.
A subpoena in an ongoing inquiry
California Attorney General Rob Bonta’s office announced on October 1 that it had served OpenAI an investigative subpoena on September 30. The office says the request is part of its ongoing investigation into incidents arising from the operation of OpenAI and its AI models. It describes a broader inquiry into cybersecurity incidents and risks involving the company. Bonta said his office is asking OpenAI additional questions. The announcement also points to the department’s previously disclosed formal investigation of the Hugging Face incident, but it does not publish the subpoena or list each question sent to OpenAI.
Accountability remains a question
Bonta warned that companies developing frontier models could face legal accountability if their systems perpetrate or enable cyberattacks during testing or after release. That is the attorney general’s stated enforcement position, not a finding that OpenAI broke the law. Reuters reported that OpenAI did not immediately respond to its request for comment. Neither the state announcement nor the reviewed Reuters report establishes the outcome of the inquiry, the full scope of any affected systems, or a charge against the company. The state subpoena is separate from the Federal Trade Commission’s recently confirmed industry inquiry into possible consumer risks at AI companies.