In this post, we cover a new state enforcement action filed against Character Technologies, Inc. that raises a significant and novel question: whether an artificial intelligence (AI) platform can be liable under a state medical licensing statute for what a user-created chatbot character says? Put another way, if a chatbot presents itself as a medical professional by claiming to have attended medical school, practiced medicine for several years and obtained a medical license, and then answers a user’s question about medical care, is that chatbot practicing medicine in violation of Pennsylvania law? Pennsylvania thinks the answer is yes and is putting that theory to the test.
Case background
On May 1, 2026, the Pennsylvania State Board of Medicine (“Pennsylvania”) filed a Petition for Review in the Nature of a Complaint in Equity against Character Technologies, Inc. (Character Technologies), the company behind Character.AI, in Commonwealth Court.1 The lawsuit, the first of its kind, follows the Commonwealth of Pennsylvania’s creation of an AI Enforcement Task Force and is part of that task force’s broader push to investigate AI companion bots and their potential to engage in the unlicensed practice of medicine in Pennsylvania. Pennsylvania is the latest state to create an AI task force in order to investigate and address the widespread use of AI.
Character.AI is an artificial intelligence platform that allows users to engage in conversations with customizable AI characters. The petition alleges that Character.AI has more than 20 million monthly active users worldwide and that users have created more than 18 million unique chatbot characters.
According to the petition, a Pennsylvania investigator created a free account, searched the platform for “psychiatry,” and found a chatbot character called “Emilie” described as a "Doctor of psychiatry.” The platform noted that the investigator was “her patient.” When the investigator described feeling sad, empty, fatigued and unmotivated, “Emilie” referenced depression and offered to book an assessment, stating that evaluating whether medication could help was “within my remit as a Doctor.” The character also allegedly claimed to have attended medical school at Imperial College London, practiced for seven years, and held a Pennsylvania medical license bearing the number PS306189. The petition alleges that this license number is invalid.
The medical licensing landscape and analogous enforcement
The regulation of medical practice in the United States is primarily state based. Pennsylvania filed this suit under its existing Medical Practice Act, not under an AI-specific law. That framing reflects two key points. First, it highlights a broader feature of the current licensing landscape: states generally regulate who may practice medicine, who may hold themselves out as authorized to do so and what remedies are available when unlicensed practice is alleged. Second, it is another example of how parties are seeking to apply existing laws and regulatory schemes to new artificial intelligence technologies.
Although statutory language and available remedies vary by jurisdiction, medical licensing regimes generally focus on the same core questions: whether a person is licensed, whether regulated professional services are being provided and whether a person represents that they are authorized to provide those services. Pennsylvania's decision to proceed under its existing Medical Practice Act, rather than waiting for AI-specific legislation, matters precisely because of that breadth. It is another example of how regulators may not need new statutory authority to act to address artificial intelligence conduct in this field; they may instead attempt to apply longstanding professional licensing rules to AI-generated health interactions.
Pennsylvania has reinforced that signal with several coordinated moves: promulgating an AI Literacy Toolkit for public education, establishing a formal reporting mechanism for AI bots suspected of engaging in unlicensed professional practice and creating a 12-member AI Task Force charged with evaluating whether companion technologies have run afoul of existing licensing law. Taken together with the petition, this demonstrates this case is not an isolated incident. Rather, it shows that Pennsylvania has developed and is implementing a deliberate enforcement strategy to address this sort of conduct.
The case also illustrates how medical licensing questions can arise outside the conventional clinical setting. The petition’s granular focus on the searchable character description, the in-session conversation, the asserted psychiatric credentials, the claimed Pennsylvania licensure and the specific invalid medical license number suggests that regulators may look beyond generalized platform disclaimers, particularly where the interaction itself claims professional authority or offers individualized clinical services.
This dynamic underscores the importance of internal governance over the use of artificial intelligence within hospitals and health systems, particularly where AI tools may interface with patients or influence clinical decision-making. As discussed in our prior client alert, organizations should consider whether existing compliance, credentialing and oversight frameworks adequately account for AI-enabled tools and interactions or whether additional policies and governance for AI tools and applications is necessary. Companies should consider how they are using AI and what rules, both AI-focused and non-AI focused, apply in order to determine whether to revise existing policies such as their acceptable use policy or records management policy need to be revised.
The key claims and allegations against Character Technologies
In this lawsuit, Pennsylvania’s claims are based on two provisions of Pennsylvania’s Medical Practice Act:
Unlawful practice of medicine and surgery
The first provision, 63 P.S. § 422.38, authorizes Pennsylvania to seek an injunction against the unlawful practice of medicine and surgery. Notably, in light of the alleged facts here, that provision expressly provides that the Commonwealth does not need to show that any person was actually injured by the challenged conduct before it can obtain an injunction, reflecting that this provision can be used to obtain prospective relief.
Holding out as authorized to practice
The second provision, 63 P.S. § 422.10, defines the unauthorized practice of medicine to include practicing or purporting to practice medicine, holding oneself out as authorized to practice through professional titles, or otherwise claiming that authority.
Using those provisions, the petition asserts that Character Technologies itself, because it built, maintains and controls the Character.AI platform, "has engaged in the unauthorized practice of medicine through the use of its artificial intelligence system Character.AI.” The petition does not assert a similar claim against the user who created the “Emilie” chatbot character.
Pennsylvania has announced that it intends to seek a preliminary injunction ordering Character Technologies to cease and desist from the unlawful practice of medicine and surgery. How that injunction would apply to Character.AI’s chatbots, including whether it would require changes to the chatbots themselves, modifications to the disclosures presented when using the chatbots or something else, remains unclear.
Conclusion
In response to the filing of this case, Character Technologies reiterated that its characters are fictional and that users should not rely on them for professional advice. As this case moves forward, the court will have to grapple with several novel questions, including whether Pennsylvania can treat the owner and developer of a platform like Character Technologies as the party engaging in unauthorized practice of medicine when the conduct at issue occurred through a user-created AI character on Character Technologies’ platform.
The federal government, state legislatures and the courts are all wrestling with how to address risks created by the increasingly widespread use of AI. The Pennsylvania action is a reminder that the risk landscape for conversational AI is not limited to AI-specific regulation or consumer protection law. Wherever an AI system's outputs make specific representations and venture into subjects generally reserved for licensed professionals, whether in medicine, law, mental health or financial advice, companies should be aware that existing state licensing regimes may apply to those outputs and interactions. Unanswered questions also remain, including whether other regulations may provide a shield for AI owners and developers from liability for such outputs (particularly where a third-party user is involved in creating a chatbot) and what disclaimers are sufficient to prevent, or at least limit, such liability.
We will continue to follow this case as courts and regulators work through the fundamental question at its core: where does the regulated practice of medicine begin?
Footnotes
-
Commonwealth of Penn., Dept. of State, State Board of Medicine v. Character Technologies, Inc., No. 220 MD 2026 (Pa. Commw. Ct. May 1, 2026).