Table of Contents
Ready to See Results?
From strategy through execution, Atlantic Health Strategies integrates compliance, operations, and growth into durable, measurable results. Let’s put our expertise to work for your organization.
The direct answer: 47 states, 250+ bills, 33 enacted
If you run a behavioral health program in more than one state, you are almost certainly out of compliance somewhere on AI right now. Lawmakers in 47 states introduced more than 250 bills regulating AI in healthcare in 2025, with 33 signed into law across 21 states, per a December 16 Manatt Health analysis published by Becker’s Hospital Review.
The pace did not slow. Manatt’s tracker reports that 43 states have introduced over 240 bills in the first months of 2026, almost as many as introduced in all of 2025. Illinois violations of the state’s new AI therapy statute carry civil penalties of up to $10,000 per confirmed violation, assessed by the Illinois Department of Financial and Professional Regulation.
There is no federal preemption to lean on. Operators have to build to the strictest applicable rule and document it.
I run Atlantic Health Strategies. My team does licensure, accreditation, and compliance work for treatment programs in states like Florida, Texas, Colorado, Tennessee, and Arizona. The AI conversation used to sit inside IT. It is now a surveyor question, a payer audit question, and a state attorney general question. If you cannot produce your AI governance file on demand, you have a finding waiting to happen.
Where the laws actually bite: chatbots, UM, and clinical decision support
Three buckets matter for behavioral health right now.
1. AI chatbots and “virtual therapists.” Illinois moved first. Governor Pritzker signed HB 1806, the Wellness and Oversight for Psychological Resources Act, in August 2025. The Act prohibits anyone from using AI to provide mental health and therapeutic decision-making, while allowing the use of AI for administrative and supplementary support services for licensed behavioral health professionals. IDFPR enforces it, with fines of up to $10,000 per confirmed violation.
Kyle Hillman, NASW-Illinois Director of Legislative Affairs, said in the IDFPR release: “With HB1806, Illinois sets the standard for regulating the use of AI in mental health.” The statute defines consent narrowly. A general terms-of-use click does not satisfy it. Some app vendors have quietly pulled out of Illinois rather than absorb the exposure.
2. Payer use of AI in utilization management. States are quickly writing rules that force health plans to keep humans in the loop. If you bill commercial or Medicaid managed care in behavioral health, an algorithm is very likely evaluating your auths. Illinois’ new law also explicitly permits AI only for administrative and supplementary support tasks, with significant restrictions: AI cannot make independent therapeutic decisions, directly interact with clients, generate treatment plans without professional review, or detect emotions.
3. High-risk AI and bias oversight. Colorado’s SB 24-205 remains the country’s most ambitious comprehensive state AI law. Governor Polis signed SB 25B-004 on August 28, 2025 to postpone the effective date from February 1, 2026 to June 30, 2026. If you deploy AI in Colorado, do not assume the current statute is what you will be complying with on day one. Track it monthly.
Why this matters for behavioral health operators specifically
Behavioral health uses AI in places most leaders do not even inventory: ambient scribes inside the EMR, predictive risk scoring on crisis lines, chat-based triage on the website, payer-facing prior authorization automation, and recovery-coaching apps that integrate with the chart. Each one touches a state rule somewhere.
Accreditors themselves are moving. Manatt noted in its Q3 update that Q3 2025 saw AI-focused accreditation tracks from URAC, guidance from the Joint Commission and the Coalition for Health AI (CHAI), and an AI workgroup convened by the National Committee for Quality Assurance (NCQA). A state surveyor, a CARF reviewer, or a Joint Commission auditor will apply that same framing when they walk in and ask what your AI governance looks like.
If your tech stack includes a chatbot vendor and you do not know whether they still operate in the states where you hold a license, that is your homework this week.
What changes when federal policy moves
Do not assume Washington will rescue you from the patchwork. Executive Order 14365 directed DOJ to establish an AI Litigation Task Force to challenge “onerous” state AI laws and instructed the Secretary of Commerce to publish an evaluation of such laws by March 11, 2026. Manatt noted, as of its most recent tracker update, that such a list has not materialized. State activity accelerated anyway.
State insurance regulators are pushing back hard. The NAIC’s Model Bulletin on insurer use of AI is being operationalized through the AI Systems Evaluation Tool, and as of March 2026 that Tool is being piloted by 12 participating states, with adoption anticipated at the 2026 Fall National Meeting. The Mayer Brown NAIC Spring 2026 summary confirms that 24 states and D.C. Have adopted the Model Bulletin, and four states have adopted insurance-specific regulation or guidance regarding AI.
For operators, the translation is short. Your state rules stand. Build governance to the strictest applicable jurisdiction. Document everything in a form a state AG, a payer SIU, or an accreditor would accept.
The AHS playbook: what to do in the next 60 days
Here is what my team walks clients through.
- Inventory every AI touchpoint. Scribes, scheduling, triage chat, predictive risk, UM and auth automation, marketing chatbots, voice agents on the call center. Name the vendor, the workflow, the data flow, and the states it touches. No inventory, no defense.
- Map each tool to the strictest applicable state rule. If you operate in Illinois, the Act bars AI from making independent therapeutic decisions, directly interacting with clients in any form of therapeutic communication, or generating therapeutic recommendations or treatment plans without the review and approval by a licensed professional. If you operate in Colorado, prepare now for the June 30, 2026 effective date of SB 24-205 and any last-minute revisions.
- Get written client consent where required. Illinois requires clear, explicit affirmative written consent for AI-supported recording or transcription. A general terms-of-service click does not count. Build a standalone consent form and train clinical staff on when it is deployed.
- Put governance in the compliance committee, not in IT. Your compliance team should own bias testing, vendor due diligence, model change logs, and impact assessments on the same calendar as your mock survey and your payer readiness review. Payors are pushing that same expectation downstream to you through contracts.
- Watch the chatbot vendors closest. Manatt reports that in the first quarter of 2026, 36 states introduced over 70 bills regulating AI chatbots, the majority of which include requirements to disclose to the chatbot user that they are interacting with an AI chatbot, not a human. Your marketing team’s website widget can create licensure exposure your clinical leadership never authorized.
None of this requires a six-figure consulting engagement. It requires an owner. Pick one. Give them the calendar, the budget, and the authority to say no to a vendor.
Frequently asked questions
How many states passed AI healthcare laws in 2025?
Manatt Health’s tracker documents that 47 states introduced more than 250 healthcare AI bills in 2025, and 33 were signed into law across 21 states, per a December 16, 2025 Manatt analysis reported by Becker’s Hospital Review. State activity accelerated in 2026, with 43 states introducing over 240 additional bills in the first months of the year, and 36 states introducing over 70 chatbot-specific bills in Q1 2026 alone.
Does Illinois’s WOPR Act ban all AI use in behavioral health?
No. HB 1806, signed August 4, 2025, prohibits AI from providing mental health and therapeutic decision-making and bars AI from directly interacting with clients in therapeutic communication, generating treatment plans without licensed review, or detecting emotions. It permits AI for administrative support (scheduling, billing) and supplementary support when a licensed professional reviews outputs. Enforcement sits with the Illinois Department of Financial and Professional Regulation, which can impose fines of up to $10,000 per confirmed violation.
How many states have adopted the NAIC Model Bulletin on insurer use of AI?
According to the NAIC’s April 2026 Spring National Meeting summary reported by Mayer Brown, 24 states and D.C. Have adopted the Model Bulletin, and four additional states have adopted insurance-specific AI regulation or guidance. The NAIC’s AI Systems Evaluation Tool is being piloted by 12 participating states from March through September 2026, with adoption anticipated at the 2026 Fall National Meeting. Behavioral health operators feel this pressure indirectly, through payer contracts that push governance expectations downstream.
If federal policy tries to preempt state AI laws, do operators still need to comply?
Yes. Executive Order 14365 directed DOJ to challenge state AI laws and Commerce to evaluate them by March 11, 2026, but Manatt reports that the Commerce list has not materialized and state activity has accelerated, with 43 states introducing over 240 bills in early 2026. Build to the strictest applicable state rule and adjust only if a federal standard actually displaces it.
References
- Becker’s Hospital Review: 47 states introduced healthcare AI bills in 2025 (Dec. 17, 2025)
- Manatt Health: Health AI Policy Tracker (2026 update)
- Manatt Health: Health AI Policy Tracker Q3 2025 Update
- Illinois Department of Financial and Professional Regulation: Gov. Pritzker Signs Legislation Prohibiting AI Therapy in Illinois (Aug. 4, 2025)
- Illinois General Assembly: HB 1806 Bill Status
- Colorado General Assembly: SB 24-205 Consumer Protections for Artificial Intelligence
- Akin Gump: Colorado Postpones Implementation of Colorado AI Act (Sept. 2025)
- NAIC: Artificial Intelligence topic page and AI Systems Evaluation Tool pilot
- Mayer Brown: NAIC Spring 2026 National Meeting Highlights (April 2026)