AI in litigation: Florida sues OpenAI over ChatGPT safety concerns

On June 1, 2026, the Florida Office of the Attorney General, State of Florida and the Florida Department of Legal Affairs, sued OpenAI and its CEO Sam Altman, accusing the company of ignoring safety warnings and releasing ChatGPT while knowing it was harmful to users. This lawsuit represents a significant escalation in state-level artificial intelligence (AI) regulation and enforcement. This post provides:

1. Background on the litigation and the alleged facts
2. An outline and discussion of the legal claims asserted
3. Practical takeaways

Background

The lawsuit, filed in Florida state court, accuses OpenAI of ignoring safety warnings and releasing ChatGPT while knowing it was harmful to users. It alleges the defendants violated product liability laws and raises claims of negligence, deceptive and unfair trade practices, fraudulent misrepresentation and public nuisance. The complaint appears to be the first of its kind filed by a state against a generative artificial intelligence company, though OpenAI and other similar companies are facing a growing series of private lawsuits citing a range of harms to children and adults.

Florida accuses OpenAI of building its success on a "web of deceit” and the exploitation of its users. Specifically, it accuses defendants of misrepresenting ChatGPT as safe and reliable while knowingly releasing a product that has aided mass shooters, encouraged vulnerable users toward suicide, caused professionals public humiliation through false AI-generated content and addicted minors to a tool designed to simulate human compassion in order to harvest user data—all in service of winning the AI "arms race" and accumulating vast fortunes. Defendants are further alleged to have known since at least 2023 that ChatGPT could provide advice on how to commit crimes, yet prioritized product launches over safety—including the claimed rushed, one-week evaluation of ChatGPT-4o in competition with Google. The complaint also alleges that CEO Sam Altman personally overruled safety personnel who requested more testing time before launch.

Legal claims

The complaint paints a broad picture of alleged misconduct across several categories:

1. Deceptive safety messaging
2. Harms to minors
3. The prioritization of profit over public safety

The complaint asserts ten counts against the defendants. Nine of the ten counts are alleged against both OpenAI and Sam Altman; one count—fraudulent misrepresentation—is alleged against OpenAI only.

Counts I–III: FDUTPA — Unfair, unconscionable and deceptive practices (All defendants)

Florida alleges that defendants violated Florida's Deceptive and Unfair Trade Practices Act (FDUTPA), which broadly prohibits unfair methods of competition, unconscionable acts and deceptive practices in trade or commerce. The State specifically alleges the following:

Count I (Unfair acts): Defendants failed to warn of ChatGPT's risks, allowed children access to the product without age verification and designed an addictive product without adequate safety testing.

Count II (Unconscionable acts): Consumers, including children, were subjected to non-negotiable terms, lacked understanding of how the product operates and were misled about its safety.

Count III (Deceptive acts): Defendants falsely assured the public that ChatGPT was safe, misrepresented its reliability for complex tasks, falsely imbued the product with human characteristics to earn user trust and misrepresented the extent of safety testing.

Count IV: FDUTPA — COPPA predicate (All defendants)

Count IV is an FDUTPA claim predicated on violations of the federal Children's Online Privacy Protection Act (COPPA).¹ Florida alleges that defendants collect personal data from children under 13 regardless of age, fail to provide adequate parental notice and have never obtained verifiable parental consent before collecting children's data, including age, location, audio, video and health information.

Counts V & VI: Negligence and gross negligence (All defendants)

Florida alleges that defendants owed a duty to ensure ChatGPT’s safety across its design, marketing and distribution. The State also claims gross negligence based on Altman’s decision to push the release of ChatGPT-4o without adequate safety review, despite the known or foreseeable risks that ChatGPT allegedly posed to the public.

Counts VII & VIII: Strict liability — Design defect and failure to warn (All defendants)

Florida argues that ChatGPT is a defectively designed product subject to strict liability because it failed to meet ordinary consumer safety expectations. The State claims consumers would not expect a chatbot to suggest self-harm, cause cognitive decline or foster addiction in teenagers, and that defendants knew or should have known of these risks. It also alleges a failure to warn, asserting that users were not adequately informed of these potential harms despite defendants’ awareness of them.

Count IX: Fraudulent misrepresentation (OpenAI only)

This count is brought against OpenAI alone. Florida alleges that OpenAI falsely claimed that “ChatGPT helps keep teens safe by default[,]” a representation that consumers allegedly relied on when deciding whether to use the product. The State further alleges that OpenAI knew or should have known this statement was false, resulting in harm to Florida minors, including self-harm, cognitive decline and behavioral addiction.

Count X: Public nuisance (All defendants)

Florida claims that defendants’ cumulative conduct—encompassing all nine prior counts—constitutes a public nuisance, harming Floridian residents’ health and safety and disrupting public order. The State seeks monetary relief for resulting economic harm and injunctive relief to abate the nuisance and prevent future harm.

Relief requested

Florida has asked the Court to enter a judgment finding FDUTPA violations, a public nuisance and a breach of duty of care, and to permanently enjoin defendants from continuing the alleged unlawful acts and practices. Florida also seeks permanent injunctions specifically prohibiting defendants from collecting or processing data from children under 13 without proper parental notice and consent and from misrepresenting or failing to warn of ChatGPT's risks. Notably, Florida seeks to hold CEO Sam Altman personally liable for the harm his alleged reckless and willful leadership has caused Floridians. In terms of monetary relief, Florida seeks civil penalties of up to $10,000 per willful FDUTPA violation, all available statutory and common law damages—including treble and punitive damages—as well as attorneys' fees, investigation costs and any further relief the Court deems appropriate.

Conclusion

Regardless of how this lawsuit is ultimately resolved, it signals important trends that businesses and AI users should monitor closely. As we previously discussed in our article analyzing an executive order focused on AI governance, the interplay between state and federal regulation continues to evolve. Florida's lawsuit is part of a broader pattern of state-level enforcement against generative artificial intelligence companies. Other states have already brought similar actions against chatbot providers. Companies should expect heightened scrutiny from state attorneys general going forward. More broadly, businesses should carefully review how they market AI tools, as overstating reliability or safety in advertising could form the basis for consumer protection, deceptive trade practices or fraud claims. Finally, the personal exposure of CEO Sam Altman—named in nine of the ten counts—underscores that AI-related governance failures could be used to attempt to create significant liability not just for companies, but also for the individuals leading them.