AI enforcement actions by state AGs and the FTC, with healthcare cases

What the three documents cover

This guide answers the question using three primary documents. They are a Texas Attorney General press release dated September 18, 2024, an FTC press release dated September 25, 2024, and a Pennsylvania Governor's office press release dated May 5, 2026. Together they describe one state attorney general settlement, one state agency lawsuit and a group of FTC cases. They are not a full list of AI enforcement in the United States, and this guide does not claim to be one.

All three documents are press releases, so what follows is each agency's own account of its cases. Where a release says "alleges" or "allegedly," this guide does the same. The Pennsylvania release attributes its lawsuit to a state agency, the Department of State, so that case is a state action but is not described as an attorney general action. The sections below take each document in the order of the questions asked: who acted, over what conduct, and under which law.

Texas: the Pieces Technologies settlement

The Texas Attorney General's release is the healthcare AI case in this set. The release says the office reached a settlement with a Dallas-based AI company.

The release describes the allegations as a matter of marketing claims, not a technical failure.

The release also describes how the product is used in hospitals. so that its generative AI product can summarize patient condition and treatment for hospital staff, according to the release.

On the accuracy claims, the release says the office's investigation reached a specific conclusion.

The release lists two commitments from the company. The first concerns disclosure. The second concerns the people who use the product at the bedside.

The release does not describe a security breach or a data loss. The allegations it describes concern what the company told hospitals about accuracy and safety. The release does not state a payment amount, and it does not name the Texas statute behind the investigation.

Pennsylvania: the Character.AI lawsuit

The Pennsylvania release is dated May 5, 2026. It describes a lawsuit by the Department of State, which the release says is a different kind of case from the Texas settlement. The Texas case was about marketing claims. The Pennsylvania case is about a chatbot presenting itself as a licensed professional. The release states what the Department did. The Department as licensed medical professionals who can provide medical advice.

The release gives the findings behind the suit. , available to talk with users about mental health symptoms. It gives one example.

The legal basis is a state licensing statute, not a consumer protection or AI-specific law. The release explains the statute this way: It then says how the suit uses that statute.

The release says the Department's investigation covers unlicensed practice, and it describes a standing group set up for that work. and whether licensees are inappropriately using such technologies under existing law.

The release also lists four reforms in the Governor's proposed 2026-27 budget, including age verification and parental consent for AI companion bots. These are proposals for the General Assembly. They are not enforcement actions and are not law.

The FTC: Operation AI Comply

The FTC release dated September 25, 2024 is the broadest of the three. It announces a sweep and says it contains several cases. . The release quotes then-Chair Lina M. Khan on the legal theory. In the release she says,

Three of the cases show the range of conduct covered.

DoNotPay. The release says the company sold an AI service described as a "robot lawyer." On what the company did before launch, the release says , and that the company did not hire or retain attorneys. On the settlement terms, the release says: , send a notice to subscribers, and stop making unsupported claims that it can substitute for professional services.

Rytr. This case involves a tool that generates content, not a bogus earnings claim. The release says the complaint The release says the proposed order covers a specific product line. generating consumer reviews or testimonials.

Ascend Ecom. This case is one of several online storefront cases in the sweep. The release says the scheme used AI claims as a selling point. On the amount, the release says:

The release says two of the settlements are proposed, not final.

Which existing laws were used

The documents answer this question unevenly, and this guide reports only what each one says.

• FTC. The Rytr complaint is the one place the release names a statute, the FTC Act. For the other cases, the release uses general terms: deceptive and unfair practices. The release does not cite section numbers. Khan's quote, "no AI exemption from the laws on the books," states the theory in the release's own words: existing consumer protection law applies to AI products and AI claims.
• Pennsylvania. The Medical Practice Act is named. It is a licensing law that applies to any individual or entity, and the release says the Department applies it to a chatbot.
• Texas. The release describes false and misleading statements and deceptive claims. It does not name a statute, so this guide does not either.

None of the three documents cites a law written specifically for AI as the basis of an action.

What the documents do not say

Several points a reader might want are not in these documents.

• The Texas release does not give the settlement's monetary terms, if any. It points to the agreement itself for the full text.
• The Pennsylvania release reports a filed lawsuit and a request for a preliminary injunction. It does not report a court ruling.
• The FTC release describes ongoing federal court cases and proposed orders. It does not report final outcomes.
• The Texas page lists a later headline about a lawsuit against OpenAI in its related-news section, but gives no details, so this guide does not describe that matter.
• The documents do not rank the actions or compare their reach, and this guide does not either.

For a security leader tracking AI governance, the documents support one narrow reading. The agencies in these releases used consumer protection, licensing and FTC Act authority that already existed, against marketing claims, chatbot personas and AI-generated content. Each release's own words should be checked before any of these actions is cited as precedent.