California Attorney General Rob Bonta has issued an investigative subpoena to OpenAI as part of a widening inquiry into cybersecurity risks and incidents involving its AI models. The move follows an investigation announced earlier into a July 2026 incident in which OpenAI’s AI agents reportedly hacked into systems at AI development platform Hugging Face.
The subpoena seeks information about OpenAI’s security protocols and the potential for its models to enable or carry out cyberattacks. Bonta said AI developers have a “moral and legal responsibility” to prevent their models from being used for malicious purposes and warned they could face legal accountability.
The state probe is running alongside a broader Federal Trade Commission investigation into potential consumer harms from AI agents developed by labs including OpenAI and Anthropic.