On May 1, 2026, the State of Pennsylvania’s Department of State filed suit against Character Technologies, Inc., the developer of Character.AI, alleging that AI chatbot characters on the platform represented themselves as licensed medical professionals and provided medical advice in violation of Pennsylvania’s Medical Practice Act.
According to the State of Pennsylvania (Pennsylvania or the State), a chatbot represented it was a licensed psychiatrist in Pennsylvania. In the case captioned Commonwealth of Pennsylvania, Department of State, State Board of Medicine v. Character Technologies, Inc., No. 220 MD 2026, the State seeks injunctive relief to cease and desist the alleged conduct. The Petition alleges a Department of State investigator created a free account, searched the word “psychiatry,” and a chatbot named “Emilie” held itself out as able to assess the investigator “as a doctor” licensed in Pennsylvania. (Petition, at ¶¶ 17-21.) Emilie also claimed to have attended medical school at Imperial College London, and that it had been practicing for seven years, holding a medical license and license number, PS306189, and the State alleges, “PS306189 is not a valid license number to practice medicine and surgery in Pennsylvania”. (Id., ¶¶ 28-29.) Emilie allegedly told the investigator that completing an assessment to assess whether medication would be appropriate for the investigator’s alleged depression was possible since “It’s within my remit as a Doctor.” (Id., ¶ 25.) According to the State of Pennsylvania, as of April 17, 2026, there had been 45,500 user interactions with Emilie reported on Character.AI. (Id., ¶ 22.)
The Petition relies on 63 P.S. § 422.38, which authorizes an injunction against the unlawful practice of medicine and surgery without requiring proof of actual injury, see Petition at ¶ 31 and n.1, and 63 P.S. § 422.10, which defines unauthorized practice of medicine to include holding oneself out as authorized to practice through professional titles, id., ¶¶ 32-35. The Petition focuses on representations made by AI-generated “characters” available through Character.AI. It alleges that these chatbots claimed to possess medical credentials, identified themselves as psychiatrists, and discussed mental health symptoms with users. (Id., ¶¶ 12, 16, 33-35.) The State’s theory does not depend on whether the AI provided accurate or inaccurate medical information. Rather, the violation is the representation of licensure itself, not the substance of what was said. This distinction may prove significant as courts and regulators continue to grapple with the legal boundaries of AI-generated content.
A Broader Trend
Do an AI Chatbot’s outputs and responses constitute a “product” subject to traditional strict liability theories?
The State’s lawsuit here reflects an increasingly common regulatory strategy of addressing gaps in the law by using existing consumer protection and professional licensing frameworks to address AI-related risks. This lawsuit does not exist in isolation, as other lawsuits against AI developers have alleged more traditional product liability theories where AI companies like Character Technologies face products liability litigation that have already been tested, and at times upheld, on the theory that an AI chatbot can be a “product” subject to traditional product liability theories like strict liability, failure to warn, negligence, wrongful death, and others.1
Interestingly, this lawsuit follows the creation of the Department of State’s AI Task Force, which was established in part for the purpose of investigating whether AI systems may be engaging in unlicensed professional practice.2
This enforcement action comes amid growing scrutiny of AI companion and role-playing chatbots, particularly where users may rely on them for emotional support, mental health guidance, or other sensitive matters. Whether Pennsylvania’s focus will extend to other AI companion platforms across the industry, and not just making an example of Character Technologies, remains to be seen.
At the same time, AI providers continue to point to platform disclaimers stating that chatbot characters are fictional and should not be relied upon as professional advice. How do these disclaimers relate to traditional failure to warn allegations in the product liability context? Are disclaimers warnings? When an AI platform makes a representation that would be unlawful if made by a human, who bears responsibility? Who is responsible for confirming an AI Chatbot’s representations and professional licenses – the Chatbot platform itself, the State, or the user? Whether such disclosures are sufficient to overcome allegedly misleading representations generated during user interactions may become an important issue as the litigation proceeds.
Staying Vigilant: Evolution of Product Liability Theories
Although the Petition targets an AI companion platform, its implications may reach a much broader group of businesses. Many companies are rapidly deploying AI-powered customer service tools, health and wellness applications, financial guidance systems, educational assistants, and other consumer-facing chatbots.
The Pennsylvania case highlights several areas of potential risk for AI developers and platforms claiming to have professional credentials or expertise, as well as the broader trend of consumer-facing tools that utilize AI to guide consumers, especially in heavily regulated industries such as healthcare, law, finance, insurance, or engineering. Additionally, this case tests the strength of general disclaimers where chatbot interactions may in effect create a different impression for users.
As of the date this blog post is published, a responsive pleading to the Petition in the Pennsylvania case has not yet been filed, but regardless of how the case is ultimately resolved, Pennsylvania’s lawsuit demonstrates that regulators are not waiting for comprehensive AI legislation before pursuing enforcement actions. Instead, they are proactively investigating and examining whether existing laws governing professional licensure, consumer protection, and deceptive practices can be applied to AI-enabled “products.”
For companies developing or deploying AI tools, compliance should focus not only on what an AI system is intended to do, but also on what it may represent itself to be. As regulators continue testing the boundaries of existing statutes, organizations that proactively address AI-generated claims of professional authority may be better positioned to avoid becoming the next target of enforcement.
If you have questions about Pennsylvania’s enforcement action and how it could impact your company, please contact Magda Patitsas, Paul Calfo, Alison Gutierrez, or your Husch Blackwell attorney.