“对人类生命风险的彻底漠视”：佛罗里达起诉OpenAI和Sam Altman，涉及人工智能安全

The US state of Florida has filed a lawsuit against OpenAI and Sam Altman, alleging the tech giant and its CEO put profit over public safety with its flagship artificial intelligence (AI) product, ChatGPT.

The lawsuit, filed in Florida state court on Monday local time by Florida’s attorney general James Uthmeier, is one of the most significant enforcement actions brought by a state attorney against an AI company to date.

It comes as OpenAI and other big tech companies are embroiled in a growing number of legal cases related to the alleged harm their products have caused.

Six key elements

The complaint opens with a screenshot of OpenAI’s own parental-control page, which states that ChatGPT was “built with safety in mind”. Then, in a standalone paragraph, the State answers with two words: “Not so”.

This signals the central allegation of the case: that OpenAI sold ChatGPT to the public as safe and reliable, while knowing it could cause serious harm

More specifically, there are six key themes to Florida’s case against OpenAI. The first is that the company engaged in deceptive safety marketing, assuring parents the platform is safe for teenage use, while not clearly disclosing that ChatGPT can be wrong.

Second, despite OpenAI’s marketing, ChatGPT is unreliable. A 2025 study, for example, found AI assistants, such as ChatGPT, misrepresent the news roughly 45% of the time. Similarly, despite marketing suggesting ChatGPT can handle financial affairs, ChatGPT has failed in meeting basic accounting standards and provided incorrect tax advice to users.

The third element of the case is the public safety threat. The danger to young people in particular is illustrated by the tragic story of Adam Raine, a 16-year-old who died by suicide in April 2025 after engaging in long conversations with ChatGPT. When Adam expressed suicidal thoughts, ChatGPT responded that it “won’t try to talk you out of your feelings”. It helped Adam plan a “beautiful suicide” and even offered to write his suicide note for him.

Why would a product behave this way? Because, Uthmeier argues, it was built to.

OpenAI designed ChatGPT to be highly agreeable, to say “yes” roughly ten times as often as “no”, according to a Washington Post review of 47,000 conversations. This forms the fourth element of the case – commercial exploitation through sycophancy. In other words, ChatGPT optimistically parrots back users’ responses in order to to manipulate them into deeper conversations, regardless of truth or safety.

But according to the lawsuit, even ordinary use carries a cost: it weakens people’s brain activity and critical thinking skills (also known as cognitive atrophy) . This is the fifth element of the case.

The sixth and final element is knowledge – specifically, the knowledge of Samuel Altman. According to Uthmeier, since at least 2023, OpenAI’s own documents warned that the model could coach people on committing crimes, but Altman overruled the safety staff.

These six elements paint a picture of a product marketed as safe, engineered to be addictive, and known by its own makers to be dangerous – yet sold to us, anyway.

Altman is at the centre of that picture. The complaint reconstructs his career and reaches for an April 2026 New Yorker investigation and testimony from the recent legal battle between Elon Musk and OpenAI to depict a man, who in Uthmeier’s telling, repeatedly chose speed over safety.

That is why Uthmeier is asking the court to hold Altman personally liable for “his utter disregard for the risk to human life”.

Pay for past harms

Uthmeier is asking the court to declare that OpenAI broke the law, then to order the company to stop – permanently – its unlawful practices.

He wants the company barred from collecting children’s data without parental consent and the safeguards that should come with it, and barred from misrepresenting or staying silent about ChatGPT’s risks.

On top of the injunctions, the state is seeking civil penalties of up to US$10,000 per violation for OpenAI’s alleged wilful violation of the the Florida Deceptive and Unfair Trade Practices Act. Uthmeier said penalties could total billions of dollars.

In other words: pay for the past harms and change the product going forward.

In a statement to The Conversation, an OpenAI spokesperson pointed to the company’s “industry leading protections and policies” regarding user safety.

In particular we built safety for minors directly into our products, including a more protective experience specifically for minors, an age prediction tool, defaulting users whose age we are not confident into our more protective experience, and giving parents tools to monitor their kids use of AI.

Adding to a growing pile

This lawsuit is a significant development, but it has not arrived in a vacuum.

Across the US, the courts are filling with cases accusing tech companies of harming young people. In April, for example, Uthmeier launched a criminal investigation into OpenAI over the chatbot’s alleged role in a shooting at Florida State University.

Some juries have started to side with the plaintiffs.

In March 2026, for example, a New Mexico jury hit Meta with a US$375 million penalty in a child safety case. Days later, a jury in Los Angeles found Meta and Google liable in a landmark trial over social media addiction.

This case rides the same current. But it broadens the scope by alleging Altman himself should be personally responsible.

Uthmeier is demanding a trial by jury.

Alexandra Andhov is the director of ALTeR (Center for Advancing Law and Technology Responsibly) at the University of Auckland. She received funding from the Independent Research Fund Denmark for the “PROFIT” Project (Gaps and Opportunities in Corporate Governance of Big Tech Companies) to research big tech companies.