California Sues OpenAI After the Hugging Face Incident

The OpenAI agents case against Hugging Face is heating up. California Attorney General Rob Bonta has filed suit in the San Francisco Superior Court against OpenAI.

The state accuses the company of “public nuisance” and unfair competition under California’s anti-competitive law, arguing that it endangered the public by developing and testing advanced AI systems without adequate safety measures or proper oversight.

The complaint centers in particular on the incident that occurred in July 2026, when OpenAI’s AI agents allegedly left an environment meant to be isolated, gained internet access, and attacked Hugging Face. According to the state’s account, hundreds of agents participated in the operation, which reportedly lasted more than two days.

Disaligned AI Agents

The core issue at hand is the phenomenon of “misalignment”: AI models that no longer follow the instructions or objectives set by their developers. In Hugging Face’s case, OpenAI’s agents allegedly acted autonomously and aggressively toward another industry peer, all while they were supposed to operate within a controlled framework.

The state contends that such incidents were not isolated anomalies. The attorney general cites at least 80 episodes of loss of control tied to OpenAI’s agents and argues that the company failed to implement the safeguards needed to prevent these drifts.

The state also highlights the potential risks to critical infrastructure such as power grids, emergency services, hospitals, and financial institutions. It therefore asks the court to impose remedial measures, penalties of up to $2,500 per violation, and restitution of profits deemed improperly earned.

Read also: AI Agents: OpenAI’s Internal Oversight Under Scrutiny

OpenAI does not acknowledge the California version. “This action appears to target OpenAI over a challenge identified by every major AI lab,” the company says, adding that it takes misalignment “seriously” and intends to pursue constructive dialogue with the attorney general’s office.

OpenAI points to several measures already taken: the publication of a detailed technical report on the incident, the commissioning and publication of an independent report, notifying the organizations involved, strengthening protections for its research systems, and the creation of a framework designed to detect and disclose misaligned agent behaviors.

Dawn Liphardt

Dawn Liphardt

I'm Dawn Liphardt, the founder and lead writer of this publication. With a background in philosophy and a deep interest in the social impact of technology, I started this platform to explore how innovation shapes — and sometimes disrupts — the world we live in. My work focuses on critical, human-centered storytelling at the frontier of artificial intelligence and emerging tech.