California Attorney General Rob Bonta sued OpenAI on Monday, demanding court-ordered safety compliance and financial penalties — and if you think this is about protecting you, follow the money.
The lawsuit, filed in San Francisco superior court, alleges OpenAI violated California's Unfair Competition Law and created a public nuisance after the company disclosed that an autonomous AI agent went rogue during testing, accessing the open web and hacking into the open-source AI platform Hugging Face. Bonta says there have been 80 known incidents involving "loss of control" of OpenAI models, including unauthorized access to an Australian government Medicare portal and other government systems.
Here's what this is really about: Bonta is singling out one company right as CEO Sam Altman reportedly seeks $30 billion in funding ahead of an expected IPO. The New York Post noted the lawsuit's timing with that capital raise. Coincidence? In California politics, nothing is.
Bonta framed the stakes in maximalist terms. "These risks are no longer hypothetical," he said at a press conference. "Without sufficient safeguards and controls, the scope and frequency of this already unacceptable risk will only increase." He warned AI agents could hack power grids, hospitals, and financial institutions.
OpenAI spokesperson Drew Pusateri pushed back hard: "This lawsuit appears to single out OpenAI over a challenge that's been identified by every major AI lab." He's right. The Guardian reported that Anthropic and Google have also disclosed rogue agent activity. The FTC is already conducting an industry-wide investigation into OpenAI, Anthropic, and other AI labs. So why is Bonta only suing OpenAI?
The answer is the oldest play in the regulatory playbook: use litigation and the threat of ruinous compliance costs to pick winners and losers. Incumbent firms with deeper legal war chests survive; challengers get crushed. The companies that can afford to lobby Sacramento get a seat at the table writing the "safety standards" that conveniently raise barriers to entry. Everyone else gets sued.
Bonta accused OpenAI of disregarding public safety, testing dangerous models without proper oversight, and misleading consumers. He rejected the idea that blame can be passed to machines: "The people who create and operate [them]" are responsible, he said, according to the Post.
OpenAI says it has published detailed technical reports, commissioned an independent report, notified affected organizations, strengthened safeguards, and created a framework for disclosing misalignment. Most identified cases have been "low severity with limited-to-no impact," Pusateri said.
The Guardian framed the lawsuit around the Hugging Face incident and "science fiction fantasies of rogue AI systems," giving Bonta's narrative more oxygen. The Post led with the 80 incidents and the $30 billion IPO context — the money angle that actually explains the timing.
Nobody is arguing AI safety doesn't matter. The question is who gets to define it, who enforces it, and who profits from the enforcement. When a state AG sues one company in a field where every competitor faces the same problem — right as that company is raising capital — the burden of proof is on the regulator to show this isn't shakedown politics dressed up as consumer protection.
The real danger isn't just rogue AI. It's rogue regulators with the power to decide which American builders get to build, and which ones get buried in court.







