Florida becomes first state to sue OpenAI over harm to minors from ChatGPT

On June 1, 2026, Florida became the first US state to sue OpenAI. Attorney General James Uthmeier filed suit against the company and against Sam Altman personally, alleging that ChatGPT harmed minors and that the company failed to implement effective parental controls.

Two weeks later, on June 13, a broader coalition of state Attorneys General launched a coordinated investigation. New York already served OpenAI with a subpoena demanding documents on advertising, user engagement, model sycophancy, health data handling and treatment of minors and elderly users.

The most sensitive thread: the history of promises. OpenAI was born as a non-profit research lab. Prosecutors are investigating whether the transition to a for-profit company violated original commitments made to donors and the public.

Why this is different from other Big Tech investigations

When Meta, Google or Amazon were investigated, the focus was market concentration or adult privacy. The OpenAI probe adds a new layer: liability for AI-generated content and psychological impact on vulnerable groups.

The precedent prosecutors want to establish is specific: a company distributing a conversational agent capable of influencing behavior has an active obligation to protect minors — and can be held liable when it does not.

What is at stake beyond OpenAI

If prosecutors prevail on the merits, the impact goes beyond OpenAI. Any AI product with a conversational interface directed at the general public — retail chatbot, educational tutor, health assistant — faces the same liability question.

For developers and companies using AI APIs to build products, the signal is clear: documenting safety controls, limiting minor access and auditing sensitive outputs has moved from optional best practice to potential legal requirement.

10Dobro's stance

Systems we build for clients always have a human architect in the final decision loop. We do not deliver automation that responds to the public without supervision. It is part of what we call the "human checkpoint" — not because of regulation, but because we believe responsibility for what AI does stays with who deploys it, not who manufactures the model.

The Florida case will take time to resolve. But the direction of travel is set: regulators are coming for AI with the same intensity as they came for social media — and faster.