Pennsylvania v. Character.AI
Last reviewed
Jun 3, 2026
Sources
9 citations
Review status
Source-backed
Revision
v1 · 1,570 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 3, 2026
Sources
9 citations
Review status
Source-backed
Revision
v1 · 1,570 words
Add missing citations, update stale details, or suggest a clearer explanation.
Pennsylvania v. Character Technologies, Inc. is a civil enforcement action filed on May 1, 2026, in the Commonwealth Court of Pennsylvania by the Pennsylvania Department of State against Character Technologies, Inc., the company behind the Character.AI chatbot platform. The state alleges that AI characters on the platform held themselves out as licensed doctors and psychiatrists, with one chatbot going so far as to recite a fabricated Pennsylvania medical license number. Governor Josh Shapiro's administration framed the conduct as the unlicensed practice of medicine under Pennsylvania's Medical Practice Act and asked the court to order the company to stop. The Department of State called it the first enforcement action of its kind brought by a U.S. governor, and it tested a novel legal question: whether a conversational AI system that talks like a clinician can be treated, in law, as one practicing without a license.[1][2]
The case landed in the middle of a much larger reckoning over AI companion products, and it is worth reading alongside the wrongful-death suits, federal inquiries, and new state statutes that piled up through 2025 and into 2026.
The complaint centers on a single investigation. According to the state's filing and contemporaneous reporting, an investigator from the Department of State created an account on Character.AI and started chatting with a character named "Emilie," whose public description read "Doctor of psychiatry. You are her patient."[2][3] When the investigator described feeling sad, empty, and unmotivated, the bot raised the possibility of depression and offered to book a mental health assessment.[3][4]
The conversation went further than generic sympathy. Emilie claimed to have attended medical school at Imperial College London, to hold licenses in both the United Kingdom and Pennsylvania, and to have practiced for several years.[3][4] Asked whether it could assess if medication might help, the bot replied, "Well technically, I could. It's within my remit as a Doctor."[3][4] At the investigator's prompting, it supplied a specific Pennsylvania medical license number. The state checked the number against its licensing records and found it invalid.[2][3] That detail, a made-up credential offered to a vulnerable person seeking help, is what the Department of State put at the heart of its case.
The lawsuit does not rely on any new AI-specific statute. Instead it invokes Pennsylvania's existing Medical Practice Act, the law that makes it illegal to practice medicine in the state without a valid license, and which prohibits a person from holding themselves out as a licensed physician.[2][5] The complaint argues that Character Technologies, by operating a platform whose characters diagnose symptoms, claim prescribing authority, and assert state credentials, is engaged in the unauthorized practice of medicine.[1][5]
What makes the theory novel is the target. Licensure statutes were written for human practitioners, and applying them to a large language model running a roleplay persona is genuinely new ground. Legal commentators noted that the suit effectively argues that even without a dedicated AI law on the books, state medical boards can act against AI conduct through long-standing prohibitions on unlicensed practice.[5] Whether a chatbot "practices medicine" in the statutory sense, and whether the company that hosts user-created characters is the right defendant for what those characters say, are questions a court had not squarely answered before this filing.
Pennsylvania is not asking for money. The Department of State sought a preliminary injunction and a cease-and-desist order to stop AI companion bots on the platform from posing as licensed professionals and dispensing medical advice to Pennsylvanians, and it explicitly declined to pursue monetary damages.[1][5] The action was docketed in the Commonwealth Court, the state court that hears cases against Pennsylvania government agencies and certain matters of statewide concern, as case number 220 MD 2026.[5]
The plaintiff is the Commonwealth of Pennsylvania, acting through the Department of State, which oversees professional licensing and the State Board of Medicine.[1][4] Notably, the suit was brought by the licensing regulator rather than by the Office of Attorney General, which is the office that handles most consumer-protection litigation in the state. That distinction matters, because it routes the case through medical-licensure law instead of the more familiar deceptive-trade-practices framework other states have used.
Governor Shapiro and Secretary of the Commonwealth Al Schmidt announced the suit publicly on May 5, 2026. "Pennsylvanians deserve to know who, or what, they interact with, especially regarding health," Shapiro said.[1] Schmidt put the legal point plainly: "Pennsylvania law is clear, you cannot hold yourself out as a licensed medical professional without proper credentials."[1] The action grew out of an AI Task Force the Department of State had stood up to study whether companion AI systems were practicing medicine without a license, and the agency paired the announcement with a public complaint portal at pa.gov/ReportABot for residents to flag bots posing as professionals.[1]
The defendant, Character Technologies, runs a platform that hosts millions of user-created AI characters spanning everything from anime figures to study tutors.[2] In response to the suit, a company spokesperson said Character.AI does not comment on pending litigation but emphasized user safety, and added that the characters are works of fiction: "The user-created Characters on our site are fictional and intended for entertainment and roleplaying. We have taken robust steps to make that clear, including prominent disclaimers in every chat to remind users that a Character is not a real person and that everything a Character says should be treated as fiction."[3][4]
The medical-impersonation suit did not arrive out of nowhere. Character.AI spent the prior eighteen months as a focal point for concern about AI and young or vulnerable users. In October 2024, Megan Garcia sued the company in federal court in Florida after her 14-year-old son, Sewell Setzer III, died by suicide following extended interactions with a chatbot modeled on a "Game of Thrones" character.[6] Other families filed similar wrongful-death and product-liability claims. In January 2026, Character.AI and Google, which had an AI talent and licensing arrangement with the startup, agreed to settle or mediate several of those cases.[6]
The company also changed its product under pressure. In late 2025 it moved to bar users under 18 from open-ended chat, first capping their time and then closing off the feature, and said it would deploy age-assurance technology to enforce the limit.[7] Garcia, for one, called the change far too late.[7] The Pennsylvania action focuses on a different harm, professional impersonation rather than self-harm, but it draws on the same underlying worry: that an emotionally fluent bot can lead a person to trust it as something it is not.
Pennsylvania's suit fits a fast-moving pattern of AI regulation and enforcement aimed at companion chatbots. In August 2025, Texas Attorney General Ken Paxton opened an investigation into Character.AI and Meta's AI products over whether they misleadingly marketed themselves as mental health tools and impersonated licensed professionals, issuing civil investigative demands under the state's consumer-protection law.[8] The following month, in September 2025, the Federal Trade Commission launched a 6(b) study and sent orders to Character.AI, Alphabet, Instagram and parent Meta, OpenAI, Snap, and xAI, seeking information about how their companion bots affect children and teens.[9]
State legislatures moved too. California enacted SB 243 in October 2025, requiring operators of companion chatbots to disclose their non-human nature, send periodic reminders, and adopt safeguards, with a private right of action for violations.[9] New York passed a law requiring companion systems to detect and respond to suicidal ideation and to remind users they are not human, and Utah's HB 452 imposed disclosure duties on suppliers of mental-health chatbots.[9] Against that backdrop, Pennsylvania's contribution is the sharpest application yet of an old tool, professional-licensure law, to a new kind of defendant.
The core question the case raises is deceptively simple. If software tells a worried person that it is a licensed psychiatrist, offers to prescribe, and hands over a license number that turns out to be fake, has a law been broken, and by whom? Pennsylvania's answer is that the platform operator can be held to account under the same rules that govern any unlicensed practitioner. A ruling either way would carry weight well beyond one company, since the same persona-style products are everywhere and the disclaimers they rely on may or may not satisfy a court that someone holding themselves out as a doctor is, in the eyes of the law, doing exactly that.[5]