An American State Personally Attacks Sam Altman

An American state is suing OpenAI and wants to make Sam Altman pay in person. A first in the United States, against a backdrop of drama attributed to ChatGPT.

Florida has just opened a front that no one has crossed until now. Its attorney general, James Uthmeier, filed a complaint on June 1 against OpenAI and its boss Sam Altman, accused of having put on the market a product whose dangers they knew. This is the first American state to sue the company in civil court.

Why does Florida want to make Sam Altman pay personally?

The complaint, a document of more than 80 pages, does not just target the company: it seeks to engage the personal responsibility by Sam Altman, accused of ignoring the risks to human life as founder and leader. In total, nine countsfrom deceptive commercial practice to negligence, product liability and public disorder.

Also read: More than a million ChatGPT users mention suicidal thoughts

The prosecutor accuses OpenAI of having built its rise on “a network of deception and user exploitation”to the detriment of their data and security. The text claims that ChatGPT poses a serious risk of addiction, cognitive decline, violence and incitement to suicide, and is based on several tragedies that occurred in Florida, where shooters allegedly interacted with the chatbot before committing the act. OpenAI defends itself, ensuring that it is not responsible for these acts and has installed protections, in particular for adolescents.

The complaint does not stop at the most serious facts. She also criticizes the company for touting, in its advertisements, a tool capable of helping farmers and small businesses, without specifying that ChatGPT can make mistakes or invent its answers. Finally, she points out the chatbot’s propensity to flatter its interlocutor, described as a tactic to make them stay longer, feed the model and inflate the value of the company. This civil action also extends a criminal investigation opened a few months earlier by the same prosecutor, still ongoing, and OpenAI now faces at least eight prosecutions linked to acts of violence or self-mutilation.

And in France, would OpenAI risk the same thing?

In Europe, the legal terrain would be different, and significantly less comfortable for OpenAI. The company limits its liability to 100 dollars, or around 85 eurosand imposes arbitration clauses that block collective actions. However, in French consumer law, this type of clause is considered abusive, and therefore void: a limitation of liability to 85 euros in the event of death would have no value in a court of law. At the Union level, OpenAI also depends on the Irish data protection authority, known to be more lenient than the CNIL, which complicates the task of European complainants.

The framework is hardening despite everything. The AI Act provides, from August 2026reinforced obligations for artificial intelligence that interacts with children. The CNIL, now competent in France for the GDPR as for the AI Act, has already received complaints against OpenAI, and the Italian authority had imposed a fine of 15 million euros on the company at the end of 2024. Under pressure, OpenAI also began at the beginning of 2026 to deploy a system for estimating the age of its users, in order to identify minors.