Florida Attorney General James Uthmeier filed a historic lawsuit against OpenAI and its CEO Sam Altman in early June 2026. This is the first case where a U.S. state is suing a language model developer over risks to children and the public. The eighty-three-page document accuses the company of negligence, unfair trade practices, and allegedly concealing the dangers that ChatGPT poses. And what's crucial — this isn't just about content moderation. It's about the product's very design. This moment marks the end of the debate about whether AI companies should bear responsibility. The new question is: how much will it cost them if they don't take it on.
Listen to this article:
83 pages of accusations: What Florida finds wrong with ChatGPT
The lawsuit filed by Florida authorities is extraordinarily detailed. The key allegation is that ChatGPT is a platform where tips on eating disorders, self-harm, or even planning mass shootings are "easily accessible — including to young children." This is a fundamental difference from previous disputes — Florida is not attacking content moderation failures, but the structural properties of the language model itself. Attorney General Uthmeier claims that both OpenAI and Sam Altman personally "demonstrated complete indifference to the risk to human life." The lawsuit seeks billions in damages and a court injunction barring further operation of ChatGPT without adequate safety measures. It's an 83-page argument that legal experts say is structured similarly to successful lawsuits against the tobacco industry or social media platforms.Why now? From social media to AI
Florida is not acting in a vacuum. In March 2026, a jury in New Mexico awarded $375 million in damages in a case against Meta. The verdict: the platform deceived the public about the safety of its products and enabled the sexual exploitation of minors. Snap and TikTok opted for out-of-court settlements instead. The legal argument that opened these cases concerned algorithm design, not content moderation. And that's exactly the same principle Florida is now applying to ChatGPT. As Soren Kaplan, author of an analysis for Inc.com, writes: "Regulators have stopped asking whether companies should act responsibly. They've started calculating how much it will cost them if they don't."What this means for all companies in the AI industry
The lawsuit against OpenAI is not just one company's problem. It's a signal for the entire industry. Every company that develops or deploys AI products — from chatbots to recommendation systems to autonomous software agents — should be aware that the rules of the game are changing. Three key implications for AI companies: 1. Personal liability for leadership. The lawsuit doesn't just name OpenAI as a company, but directly targets Sam Altman. This is a warning to every CEO in the field — they can no longer hide behind corporate structures. Regulators are prepared to go after individuals. 2. Design versus moderation. If the court accepts the argument that ChatGPT's dangers stem from its architecture, it will open the door to lawsuits against any language model that wasn't designed from the ground up with safety guardrails. This means that "safety by design" ceases to be a voluntary corporate value and becomes a legal necessity. 3. Financial risk in the billions. Meta paid $375 million in a single state. Florida is seeking billions from OpenAI and Altman. If similar lawsuits spread to other states, the total financial impact could be devastating for the AI industry.European context: The EU AI Act is already in effect
While the U.S. addresses AI company liability through the courts, the European Union has taken the path of preventive regulation. The AI Act (EU AI Act), which has come into force, classifies AI systems by risk level and requires transparency, documentation, and human oversight for high-risk applications. Language models like ChatGPT fall under the rules for General Purpose AI systems. European developers and companies deploying AI must therefore reckon with stricter rules than their American counterparts. On the other hand — European companies that comply with the AI Act are more resilient to lawsuits. Compliance can become a competitive advantage. For Czech companies, this means two things: they must monitor both European regulation and the evolution of U.S. case law. Both lines will influence each other and shape standards for the entire market.OpenAI under pressure: This isn't the first lawsuit
Florida isn't OpenAI's only legal problem. In May 2026, the company faced a class-action lawsuit over allegedly sharing private user queries with Meta and Google. Before that, lawsuits emerged from families of violent crime victims where ChatGPT allegedly played a role — including the case of a Canadian mass shooter or the shooting at Florida State University (FSU). These cases create cumulative pressure that forces OpenAI — and by extension the entire industry — to reassess the balance between speed of development and safety. A company that until recently championed the "move fast" mantra must now show it can also "move safe."How AI companies should respond
Experts recommend several concrete steps:- Mandatory red-teaming — systematic testing of models for vulnerabilities before deployment, ideally by external auditors.
- Transparent documentation — detailed descriptions of training data, safety measures, and model limitations.
- Age-gating and age verification — mechanisms that limit minors' access to potentially harmful features.
- Inference-level safeguards — filters that detect and block dangerous patterns in responses in real time.
What to expect in the coming months
The Florida vs. OpenAI case will be closely watched. Its outcome will determine whether U.S. courts follow the path of precedent — similar to tobacco companies or social media. If Florida succeeds, we can expect a wave of similar lawsuits from other states — not just against OpenAI, but also against Google, Anthropic, Microsoft, and others. For regular ChatGPT users, this likely means stricter safety filters in the near future and possibly some feature restrictions for unverified users. For companies, it means one thing: the era of self-regulation in AI is over. Those who don't prepare will pay.Does the EU AI Act also apply to OpenAI's ChatGPT?
Yes. ChatGPT falls under the category of General Purpose AI systems, which are subject to specific rules of the EU AI Act — including mandatory transparency regarding training data, energy consumption, and copyright compliance systems. OpenAI must comply with these rules to operate in the EU.
Could a similar lawsuit arise in Europe or the Czech Republic?
Theoretically yes, but the mechanism is different. While the U.S. addresses liability primarily through the courts (class actions, state lawsuits), the EU prefers preventive regulation through the AI Act. However, European consumers can seek damages under existing consumer protection regulations and GDPR. The first European lawsuits against AI companies are likely only a matter of time.
What should I do as a ChatGPT user to protect my children?
Use parental controls at the device level, set age restrictions in Google/Apple accounts that limit app access. Talk to your children about the fact that an AI chatbot is not an authority and can generate harmful content. In Europe, you can also exercise your right to data access under GDPR and find out what information about your children the model processes.
