The application of AI to mental health is evolving at lightning speed. Clinicians, digital startups, and consumers alike are experimenting with AI tools that can suggest therapeutic exercises, mimic emotional responsiveness, or even give “therapy-like” advice. The state of play is both full of promise and peril.

This mismatch is creating tension, and it’s more and more essential and timely to distinguish between AI as support for therapists – like auto-notetaking or flagging risk – and AI pretending to be therapists, which are diagnosing, advising, or providing emotional support like a human would without clinical oversight. The difference between them is not semantics, consequences for public health hang in the balance. And now, regulators are stepping up with the intention to protect consumers and uphold clinical standards.

What Happened: Illinois Set a Precedent 

In Illinois, clinicians and advocates, led by the Illinois chapter of the National Association of Social Workers (NASW-Illinois), took a stand: pushing for a legal boundary on AI tools masquerading as therapists. The result: HB 1806, known as the Wellness and Oversight for Psychological Resources (WOPR) Act, which was unanimously passed in the Illinois legislature, and signed by Governor Pritzker in early August 2025. 

Illinois now stands among the very first states to place a clear legal line around AI mental health care, and the process was bipartisan. Here’s what the WOPR Act does:

• Bans AI-only therapy: AI systems cannot independently perform or advertise therapy, counseling, or psychotherapy, unless clearly tied to oversight by a licensed professional. 
• Prohibits misleading advertising claims, treating “AI therapy,” “chatbot counselor,” or “virtual psychotherapist” as misleading and unlawful unless the AI has direct clinician oversight. 
• Makes professional oversight required: Licensed mental health providers can use AI tools only if they review and approve any therapeutic output. AI cannot independently make decisions or interact with clients therapeutically.
• Allows AI for admin or support, but only under human review—tasks like scheduling, transcribing, or summarizing, with written, revocable client consent. 
• Fines up to $10,000 per violation enforced by the Department of Financial and Professional Regulation. 

Kyle Hillman, NASW-Illinois Legislative Affairs Director, put it succinctly: “At its core, the WOPR Act protects people over platforms… It ensures that mental health care is provided by trained professionals—not simulations.” 

The Ripple Effect on Other States

The passage of WOPR has influenced a new wave of proposed limitations around AI “therapists.” As of August 2025, no strong national policy on AI therapy is in sight. However, while still in early stages, state-level legislators are taking action, notably:

• New York: Senator Kristen Gonzalez introduced the NY AI Act, emphasizing the need to prevent algorithmic bias, enforce independent audits, and create a private right of action for residents. Additionally, Senate Bill S8484 has been recently introduced to impose liability for damages caused by chatbots impersonating licensed professionals – inclusive of mental health clinicians.  
• Nevada passed AB 406 in June 2025. It forbids AI systems from providing mental or behavioral healthcare or claiming they can do so. Human providers may still use AI, but only for administrative functions, and with independent review of any AI output. Violations can bring up to $15,000 fines.
• Utah passed HB 452 in March 2025, which doesn’t ban AI therapy outright but mandates safeguards: like clear disclosure that the chatbot is AI, privacy protections prohibiting data sale, and marketing restrictions. It requires mental health chatbots to clearly disclose that they are AI, not humans, and bars them from selling or sharing user data. In 2024, Utah also created an Office of AI Policy, a structural nod to the growing scope of the issue.
• New Jersey: Assembly Bill 5603 has already cleared committee, and as of June 16, 2025 is up for discussion in the New Jersey State Assembly. It bans any advertising that presents an AI system as being a licensed mental health professional. Mirroring A5603, a Senate counterpart was introduced in May 2025, and is currently in the Senate Commerce.  
• California, Texas, Oregon, Washington, Florida: Each state has proposed legislation specific to regulating the role of AI in providing mental health care, each at different points along the approval process.

Ways to Stay Up to Date on AI in Therapy Legislation

Momentum to introduce and change policy around AI “therapists” is real and growing, so if you’re looking to get involved from an advocacy perspective or checking in on what legislators in your state are doing, it’s important to stay in the know. Here are a few ways what you can do right now:

• Track the state-by-state legal landscape. Laws around AI and mental health are evolving quickly. Consider following legislation trackers to be up-to-date. Knowing about these changes arms you to advocate effectively, in keeping with professional ethics codes.
• Follow thought leaders and experts in the field. Panelists from upcoming conferences and organizations like the American Counseling Association and American Psychological Association share valuable insights on AI ethics, safety, and clinical practice.

Several states have already passed legislation restricting or banning the use of AI as a substitute for licensed therapy, with others considering similar bills. These efforts signal that lawmakers are treating the issue as a matter of public safety and professional regulation. Keeping pace with these changes helps you understand both your legal responsibilities and the direction of future policy.