Rep. Kevin Mullin, a California Democrat, introduced a federal bill on March 19 that would make it illegal for AI chatbots to claim they hold medical, legal, or financial licenses. The measure, formally designated H.R. 7985 and titled the CHATBOT Act, would treat such false claims as deceptive practices under the Federal Trade Commission Act. If enacted, the bill would create the first national standard for how chatbot providers disclose the limits of their automated advice, arriving as states from New York to California push their own versions of the same idea.
What the CHATBOT Act would actually prohibit
The bill targets two specific behaviors. First, it would bar any chatbot from falsely stating that it holds a professional license in medicine, law, or finance. Second, it would prohibit chatbots from implying that a licensed human professional has reviewed or approved the advice they generate. Both prohibitions would be enforceable by the FTC under its existing authority over unfair or deceptive acts, according to the sponsor’s official announcement.
That enforcement structure matters. Rather than creating a new regulatory body or a standalone compliance regime, H.R. 7985 plugs directly into the FTC’s existing consumer protection framework. The Commission already maintains an active docket on AI-related enforcement, including orders and information requests directed at consumer-facing chatbot providers. Federal guidance from health and consumer agencies has also emphasized that misleading representations about health services can trigger liability even when companies technically comply with privacy or data security rules. The CHATBOT Act would extend that same logic to any chatbot that dresses up its output as licensed professional counsel.
The federal bill record shows that H.R. 7985 was introduced in the 119th Congress and referred to committee, an early stage that signals interest but not yet consensus. The text is narrowly tailored: it does not attempt to regulate all AI-generated advice, nor does it set technical standards for model training or safety testing. Instead, it focuses on the specific, consumer-facing moment when a user might reasonably believe they are dealing with a licensed doctor, lawyer, or financial advisor-or with advice that a licensed person has signed off on.
State bills and teen usage patterns add pressure for a federal standard
Congress is not acting in isolation. New York’s Assembly bill A.6545 would create civil liability for damages caused when a chatbot impersonates a licensed professional. The New York proposal, as outlined in the state’s legislative text, defines key terms like “chatbot” and “proprietor,” and makes clear that pretending to be a doctor, lawyer, or other regulated professional could expose providers to lawsuits.
California’s AB-1064, titled the Leading Ethical AI Development (LEAD) for Kids Act, takes a different angle by focusing on harms from companion chatbots and social AI systems directed at minors. While its primary concern is the psychological and developmental impact of AI “friends” and tutors, it shares a core premise with H.R. 7985 and the New York bill: when AI systems present themselves as trusted authorities, lawmakers want clearer guardrails on what they can claim.
A patchwork of state rules creates obvious compliance headaches for companies that operate nationwide. A chatbot built in San Francisco, hosted on servers in Virginia, and used by a patient in Brooklyn could face overlapping and potentially conflicting disclosure requirements. For industry and regulators alike, the appeal of a federal standard is that it would set a single, enforceable floor. States could still go further-by adding private rights of action or broader duties to monitor harms-but the baseline rules about impersonating licensed professionals would be the same everywhere.
Demand-side data also reinforces the urgency, especially for younger users. A Pew Research Center report published in early 2026 documented how teens already turn to AI chatbots for schoolwork and personal questions, often in private settings where adults and educators cannot easily see what information is being provided. While that research focused on usage patterns rather than on formal legal, medical, or financial advice, it suggested that chatbots are becoming a routine part of how minors seek answers on sensitive topics. That trend heightens the stakes if chatbots blur the line between general information and what sounds like professional judgment.
Whether the bill changes chatbot marketing language
The practical test for H.R. 7985 is whether it forces chatbot providers to change what they tell users. Right now, several popular AI tools offer symptom checkers, legal document drafters, and financial planning features with disclaimers that vary widely in prominence and specificity. Some bury a “not a substitute for professional advice” line in their terms of service. Others present their output in formats that closely mimic a doctor’s note or a lawyer’s memo, complete with citations, formal headings, and confident language.
If the CHATBOT Act becomes law, providers would face FTC enforcement risk any time their interface design, marketing copy, or chatbot dialogue creates the impression of licensure. That standard is broader than simply banning a chatbot from saying “I am a doctor.” It also covers implied authority, meaning a chatbot that says “based on my review of your symptoms, you should…” could cross the line if a reasonable consumer would interpret that phrasing as coming from, or being endorsed by, a licensed clinician.
Design choices would matter as much as words. A health chatbot that uses white coats, stethoscope icons, and exam-room imagery might need to rework its visual branding to avoid suggesting that a human physician is behind every answer. Legal and financial tools that auto-fill signature blocks with professional-sounding titles could face similar scrutiny. Even the way chatbots respond to direct questions-such as “are you a real doctor?”-would likely be rewritten to provide unambiguous denials of licensure.
The likely industry response, even before any enforcement action, would be a wave of explicit disclaimers inserted into chatbot interfaces. Companies would have a strong incentive to add clear, prominent labels stating that the chatbot is not a licensed professional and that no licensed human has verified its output, at least by default. Some providers might introduce optional workflows where a licensed professional can review AI-generated drafts, but those paths would need to be clearly distinguished from automated responses to avoid misleading users.
For users, the change could be subtle but meaningful. Instead of encountering chatbots that speak in the voice of an all-knowing expert, people might see more hedged language, frequent reminders to consult a human professional, and interface cues that emphasize the system’s automated nature. Over time, consumer expectations could shift, with users learning to treat chatbot output more like a search result or reference article than a definitive professional opinion.
Open questions and what to watch next
Several gaps in the evidence limit any firm prediction about the bill’s trajectory. No public FTC complaint logs or enforcement actions specifically cite chatbot licensure misrepresentation, so the practical need for the bill rests more on observed design patterns and hypothetical risk scenarios than on a documented wave of consumer injuries. The bill text itself, as published in the federal record, contains no usage statistics or incident data quantifying how often chatbots have pretended to be licensed professionals.
That absence cuts both ways. Supporters can argue that the measure is a low-cost, preventive step that fits neatly within existing consumer protection tools. Skeptics may counter that the FTC already has broad authority to police deceptive marketing and that an AI-specific statute risks freezing design practices in place while the technology is still evolving. Lawmakers will have to decide whether a targeted rule about licensure claims is narrow enough to avoid unintended consequences yet strong enough to deter genuinely harmful impersonation.
Implementation details also remain unsettled. The bill leaves it to the FTC to interpret what counts as an “implication” that a licensed professional has reviewed or approved chatbot output. Future guidance or enforcement actions would need to clarify how far that concept reaches into user interface design, training data descriptions, and behind-the-scenes workflows where humans occasionally check AI responses.
For now, the most important developments to watch are procedural. Committee hearings, if scheduled, would reveal how lawmakers weigh the competing concerns of innovation, consumer protection, and state–federal coordination. Industry testimony could shed light on how many providers already prohibit their chatbots from claiming licensure, and whether additional federal rules would meaningfully change current practices. Until then, H.R. 7985 stands as an early attempt to answer a basic question for the AI era: who, or what, is allowed to sound like a professional-and what happens when a machine crosses that line?
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
