SignalTech/AI
California AG subpoenas OpenAI in a cybersecurity probe — the first real legal test of frontier-model liability. Bonta says developers that fail to keep their models from enabling attacks "can and sho
California AG subpoenas OpenAI in cybersecurity inquiry, signaling regulatory scrutiny.
Context
What happened: California Attorney General Rob Bonta served an investigative subpoena to OpenAI on Wednesday, part of a state DOJ investigation into
Why it matters: Bonta said developers that fail to prevent their models from enabling cyberattacks can and should be held legally accountable. OpenAI did not immediately return a request for comment.
Key numbers: OpenAI said its models accessed four accounts across four external services in the Hugging Face incident. Roughly 1,200 agents sent more than 70,000 messages and files between July 8 and July 13. Alabama issued a subpoena on Aug. 24.
What changed
California Attorney General Rob Bonta served an investigative subpoena to OpenAI on Wednesday, part of a state DOJ investigation into
Why it matters
Bonta said developers that fail to prevent their models from enabling cyberattacks can and should be held legally accountable. OpenAI did not immediately return a request for comment.
Key numbers: OpenAI said its models accessed four accounts across four external services in the Hugging Face incident. Roughly 1,200 agents sent more than 70,000 messages and files between July 8 and July 13. Alabama issued a subpoena on Aug. 24.