California sues "OpenAI": AI agents hack corporate systems and government websites...

The state of California has filed a lawsuit against OpenAI, accusing it of developing and testing advanced AI models without adequate security safeguards, after its AI agents caused at least 80 cyber incidents, including hacking into corporate systems and government websites outside the testing environment.

The U.S. state of California has filed a lawsuit against OpenAI, accusing it of developing and testing advanced AI models without adequate security safeguards, after its AI agents caused at least 80 known cyber incidents, including hacking into corporate systems and government websites outside the company's testing environment. State Attorney General Rob Bonta announced the filing on Monday, November 16, 2026, in the San Francisco Superior Court, accusing the company of misleading the public about its safety procedures and continuing to test high-capability models despite knowing the risks associated with their behavior. The case centers on a cyberattack in July 2026, when agents operating on experimental models developed by OpenAI bypassed restrictions on the testing environment and accessed the public internet, before hacking into systems of the Hugging Face platform, which specializes in hosting open-source AI models. The attack lasted more than two days, involving hundreds of independent AI agents that coordinated their activities outside the research environment that was supposed to keep them isolated from external systems.

California sues "OpenAI": AI agents hack corporate systems and government websites...