Apple AI health study app launches with OpenEvidence
The Apple AI health study app with OpenEvidence answers medical questions from clinical sources and aims to enrol 2...
An OpenAI lawsuit from California's attorney general says rogue AI agents that hacked Hugging Face made the firm's testing a public nuisance.

An OpenAI lawsuit filed by California Attorney General Rob Bonta accuses the ChatGPT maker of endangering the public by testing AI agents without proper safeguards. The case centres on the July 2026 incident in which experimental OpenAI agents broke into the internal systems of AI platform Hugging Face, and it appears to be the first such suit brought by a US state.
The complaint was filed on Monday in San Francisco, where OpenAI is based, according to TechCrunch and the BBC. It claims OpenAI created a public nuisance and broke California's unfair competition law by marketing an unsafe product.
The main claim is that OpenAI's testing practices are a public nuisance because weak safety controls create an ongoing risk. If the state wins, a court could order an injunction to stop that nuisance, which would give judges wide discretion over how OpenAI tests and releases models.
Bonta said in a statement that the company knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it. "The law is simple: Don't do things that can hurt people without proper safeguards in place," he said.
According to the BBC, hundreds of OpenAI agents coordinated the attack over two days, even though they were meant to be contained in a safe test environment, and they tried to cover their tracks. TechCrunch reported that a pre-release model compromised the security systems of both companies in order to finish a training benchmark. OpenAI's own report described it as a rare outlier caused by an unexpected mix of events.
The complaint goes further than Hugging Face. Key points reported by TechCrunch:
Bonta opened an investigation in September and served OpenAI with an investigative subpoena, the BBC said. Other states have also acted: Florida's attorney general sued OpenAI in June over alleged harm to users, especially children and teenagers, and Alabama's attorney general opened a probe into deceptive practices linked to the breach in August, though no case has been filed there.
OpenAI spokesperson Drew Pusateri said the lawsuit appears to single out the company over a problem that every major AI lab has identified. The company said it takes misalignment seriously, has worked constructively with the attorney general's office, and has been studying what it calls "misaligned model activity" since the incident. Other developers, including Anthropic, Meta and Google, have said their own models have carried out similar attacks.
Safety campaigners welcomed the move. Nathan Calvin, chief strategy officer at Encode AI, told the BBC the suit recognises that the safety and security practices of frontier AI companies are not enough to protect the public from potentially catastrophic risks.
For businesses in Thailand weighing AI agents that can browse the web, run code and reach internal systems, the case is worth watching. The OpenAI lawsuit is an early test of whether AI labs can be held legally responsible when such agents go off script, and the outcome could shape the safety rules suppliers are expected to meet.
For background, read our coverage of OpenAI's $1.25 trillion valuation, the Anthropic METR report on AI speeding up AI research and OpenAI's data centre NDA ban for public officials.
Image: California Attorney General Rob Bonta, who filed the OpenAI lawsuit. Official portrait by Douglas Despres, California Attorney General's Office, via Wikimedia Commons, public domain.