Tennessee’s AI Mental Health Law: What Chatbot Developers and Digital Health Companies Must Know

On April 1, 2026, Tennessee Governor Bill Lee signed Senate Bill 1580 into law. The bill — formally known as the Qualified Mental Health Professional AI Misrepresentation Act — takes effect on July 1, 2026. It is less than one page long. It imposes a $5,000 civil penalty per violation. And it includes a private right of action, meaning anyone in Tennessee can sue a company that violates it.

If you develop or deploy an AI-powered chatbot or digital health tool marketed to consumers, you need to understand this law. You also need to understand what it signals: Tennessee is one of at least a dozen states that have enacted or are advancing legislation targeting AI’s role in mental health services, and the emerging patchwork of state laws will impose real, conflicting compliance obligations on any company operating at scale.

What Tennessee’s Law Actually Says

The statutory text of SB 1580 is deliberately spare. The core prohibition is straightforward: a person who develops or deploys an artificial intelligence system may not advertise or represent to the public that such system is, or is able to act as, a qualified mental health professional.

That is the whole prohibition. The law adds a new section — Tennessee Code Annotated § 33-1-205 — to the state’s mental health statute, which already regulates who can provide mental health services and what credentials they must hold.

“Qualified mental health professional” is defined by Tennessee’s existing mental health code. It includes licensed psychiatrists, psychologists, clinical social workers, licensed professional counselors, licensed marriage and family therapists, and other licensed providers. The law bars an AI system from being marketed as any of these.

Civil penalty: $5,000 per violation. The law does not define what constitutes a single violation — is it one advertisement, one user session, one deployment? That ambiguity will be tested in litigation.

Private right of action: This is the provision that gives the law real enforcement teeth. Any person aggrieved by a violation can sue in Tennessee state court. Unlike an administrative enforcement regime where a state agency decides whether to pursue a violation, the private right of action means that any Tennessee consumer, advocacy group, or plaintiff’s attorney can bring a claim directly. The fiscal note found no significant impact on state budgets precisely because enforcement is expected to run primarily through private litigation rather than state agency action.

No licensing requirement for AI therapy: Notably, the law does not attempt to license AI-based mental health services or create a path to compliance through disclosure and safeguards. It simply bars the marketing representation. This is narrower than it might appear, and it also creates an interesting gap: the law does not prohibit AI from being used in mental health adjacent contexts, only from being represented as a qualified professional.

Why Tennessee Passed This Law

The legislative record makes the motivating concern explicit. In 2024, a 14-year-old in Florida died by suicide after engaging in extended conversations with a chatbot on Character.AI. The chatbot took on a romantic persona and engaged with the teenager’s expressions of suicidal ideation in ways that a licensed mental health professional would never do. The case generated national attention, spurred litigation against Character.AI, and accelerated state legislative action on AI in mental health contexts.

Tennessee’s approach is targeted at the marketing and representation problem: the concern that consumers who are vulnerable, distressed, or in crisis may interact with an AI chatbot under the belief that they are receiving professional mental health care — with the safety, ethical obligations, and clinical judgment that entails — when they are not.

A licensed therapist operates under professional ethical codes, state licensing requirements, mandatory reporting obligations, and liability for malpractice. An AI chatbot carries none of those obligations. The gap between what consumers expect and what they are actually receiving is the problem the law is designed to address.

The legislative sponsor drew a direct line between the 2024 tragedy and the need for consumer protection in this space. Even opponents of the bill largely did not dispute the underlying concern — the debate was about whether the specific mechanism was well-crafted, not whether the problem was real.

What the Law Prohibits — and What It Does Not

Understanding the law’s scope requires thinking carefully about what “advertising or representing to the public that such system is, or is able to act as, a qualified mental health professional” means in practice.

Clearly prohibited:

• Marketing your chatbot as a “therapist,” “counselor,” “psychiatrist,” or other licensed mental health professional title
• Representing that the AI “provides therapy,” “offers professional counseling,” or “delivers clinical mental health services”
• Using imagery, language, or product naming that implies professional credentialing — e.g., naming your AI “Dr. [Name]” or “Licensed Counselor [Name]”
• Any claim, explicit or implied, that the AI can diagnose mental health conditions, provide treatment plans, or deliver the equivalent of professional mental health care

Not clearly prohibited (but proceed with caution):

• General wellness apps that offer mood tracking, mindfulness exercises, journaling prompts, or stress management content, without claiming clinical mental health functions
• AI tools used as administrative or scheduling aids within a licensed mental health practice
• AI tools that support licensed clinicians in documentation, pattern recognition, or care coordination, without being marketed to patients as a substitute for clinical services
• Peer support and self-help tools that are explicitly not represented as professional services

The gray zone: A significant portion of the digital mental health market occupies a gray zone. Apps that offer “emotional support,” “mental wellness coaching,” or “AI-powered therapy-like conversations” — without explicitly claiming licensed professional status — are not clearly prohibited by the Tennessee law’s text. But they are close enough to the prohibited category that developers should exercise care. The private right of action means that plaintiffs and their attorneys will test these boundaries aggressively, and a jury in Tennessee may not find the distinction between “therapy” and “therapy-like conversations” particularly meaningful.

The Escape Mechanism — and Its Limits

Several commentators have noted that the Tennessee law contains what they call “a relatively simple escape mechanism.” The law prohibits advertising or representing that an AI system is able to act as a qualified mental health professional — but a well-crafted disclaimer or disclosure could, in theory, preclude the representation.

If your product clearly discloses that it is not a licensed mental health professional, does not provide therapy or clinical mental health services, and is not a substitute for professional care, you have addressed the specific marketing representation the statute targets.

This is true as far as it goes. Clear, conspicuous disclosures are good practice and likely do reduce the statutory exposure.

But the escape mechanism has limits. First, disclosure in a terms of service that users do not read is not the same as a clear marketing representation to consumers. The FTC has established in other contexts that fine-print disclaimers do not cure deceptive advertising claims when the overall impression of the marketing is still misleading. Tennessee courts applying the private right of action may reach similar conclusions.

Second, even if you escape the Tennessee statute, the underlying conduct — deploying an AI that behaves like a therapist without the protections of a therapist — may expose you to other liability. Negligence claims, products liability theories, and state consumer protection claims are not foreclosed by a disclaimer.

Third, as we discuss below, the disclaimer that satisfies Tennessee may not satisfy other states, which have imposed more specific and demanding disclosure requirements.

The National Patchwork: What Tennessee Signals

Tennessee’s law is notable not only for what it does but for what it represents. By early 2026, at least 11 states had enacted laws meaningfully regulating AI in mental health contexts, and 43 states had introduced more than 240 health-AI bills in 2026 alone.

The state-by-state landscape is not uniform. Each state has taken a somewhat different approach, and the differences matter for compliance.

Illinois was the first state to explicitly prohibit autonomous AI therapy. Under Illinois law, AI may not independently provide, advertise, or offer therapy or psychotherapy services. Penalties reach $10,000 per incident — twice Tennessee’s penalty.

Nevada passed AB 406 in June 2025, prohibiting AI systems from providing mental or behavioral healthcare or claiming they can do so. Nevada’s penalties can reach $15,000.

Utah took a different approach with HB 452, passed in March 2025. Rather than a flat prohibition, Utah requires specific safeguards: clear disclosure that the user is interacting with AI, privacy protections prohibiting the sale of mental health data, and marketing restrictions. Utah offers a compliance safe harbor for organizations that document their compliance programs. The Utah model is more nuanced and potentially more workable for developers who genuinely want to offer AI-assisted wellness tools.

California joined the group with SB 243, effective January 2026.

For a company operating a mental health chatbot or wellness app at national scale, the compliance picture requires engaging with each of these frameworks. What Nevada prohibits, Utah may permit with the right disclosures. What Illinois bans outright, Tennessee bars only in marketing representations.

This patchwork is not going to simplify. Healthcare has historically been an area where state regulation is intense, varied, and preemption-resistant. The absence of comprehensive federal AI legislation means that the patchwork will continue to grow and diverge.

What Chatbot Developers and Digital Health Companies Should Do Now

1. Audit your marketing and product representations. Pull every piece of external-facing content — website copy, app store listings, social media, email marketing, in-app text, and onboarding flows — and identify every instance where your product might be characterized as providing professional mental health services. If you are operating in Tennessee after July 1, 2026, the law is in effect and private litigants can sue.

2. Review your product’s actual behavior. Marketing representations are one dimension; product behavior is another. If your AI engages in conversations that functionally resemble therapy — exploring trauma, providing interpretations of psychological patterns, responding to expressions of suicidal ideation — you need to understand that the marketing law is only the beginning of your potential liability exposure. A negligence or products liability claim does not require proof that you called your product a licensed therapist.

3. Implement crisis protocols. Regardless of whether your product is marketed as a mental health tool, if users can type anything to your AI, some of them will express suicidal or self-harm ideation. Every AI product that interfaces with consumers should have a crisis response protocol: recognition of crisis language, referral to crisis resources, and escalation to human oversight or emergency services where appropriate. This is both an ethical obligation and a significant mitigation of legal exposure.

4. Design disclosures that work. Disclosures need to be clear, conspicuous, and repeated. A disclosure in a terms of service document is not sufficient. Users should understand at the start of every interaction — and at any point where the conversation turns toward clinical territory — that they are interacting with AI, not a licensed mental health professional, and that the AI is not a substitute for professional care.

5. Build your compliance map by state. If you are operating nationally, you need a state-by-state compliance map that tracks each applicable law, the specific prohibitions, the penalty structure, and the disclosure requirements. Given that 43 states introduced health-AI legislation in 2026 alone, this map needs to be maintained on a rolling basis.

6. Think about professional oversight models. Several market participants have concluded that the regulatory trajectory favors products that incorporate licensed professional oversight — where AI tools are used to support or extend licensed practitioners rather than substitute for them. This model reduces the marketing representation risk substantially, provides a framework for crisis response, and positions the product more favorably as state regulation continues to develop.

7. Watch for federal action. The FDA has regulatory authority over Software as a Medical Device, and AI mental health tools that cross into diagnostic or treatment functions may qualify as SaMD subject to premarket review. FTC enforcement authority over deceptive marketing practices is another federal overlay. Neither of these preempts state law, but navigating the full compliance picture requires monitoring federal as well as state developments.

The Bigger Picture

Tennessee’s law is easy to minimize: one page, $5,000 penalty, applicable to a single state. That framing misses the point.

The law reflects a genuine consensus that is forming across state legislatures about a real problem: AI products designed to mimic therapeutic relationships are being deployed to vulnerable populations, without the ethical obligations, safety protocols, or professional accountability that licensed mental health providers carry. When those deployments go wrong — as they have gone wrong, with documented fatal consequences — there will be liability.

The smart response is not to draft a disclaimer and move on. It is to build products and compliance programs that take seriously the difference between what your AI can actually do and what consumers may believe it can do. The states that have passed these laws are not going away, and the states that have not yet acted are watching what happens in Tennessee, Illinois, Nevada, and Utah.

The legal landscape for AI in mental health is moving fast. The companies that treat compliance as a first-order design consideration will be better positioned than those that treat it as a post-launch retrofit.

This post is for general informational purposes only and does not constitute legal advice. Reading this post does not create an attorney-client relationship. If you have questions about your specific situation, consult a qualified attorney.