States continue to set their own rules for AI in healthcare, from patient disclosure requirements and limits on AI therapy to guardrails on AI in prior authorization.
Lawmakers in 47 states introduced more than 250 bills regulating AI in healthcare in 2025, and 33 were signed into law across 21 states, according to a Manatt Health analysis.
Here are 13 healthcare AI laws to know:
1. California AB 1979: Gov. Gavin Newsom signed Assembly Bill 1979 on Sept. 30, 2026. The law ensures physicians and other licensed providers can exercise their professional judgment when AI or other clinical decision tools are used in patient care. A California Hospital Association spokesperson told Becker’s the association no longer opposes the law. Mr. Newsom vetoed a separate bill, AB 2575, that would have barred retaliation against healthcare workers who use their clinical judgment, such as by overriding AI recommendations they consider unsafe.
2. California SB 503: Mr. Newsom signed Senate Bill 503 the same day. The law requires developers of clinical decision support systems to take reasonable steps to reduce known or predictable bias. It also requires providers and others who deploy the systems to regularly monitor them and mitigate biased impacts.
3. Hawaii SB 3001: Gov. Josh Green signed SB 3001, Act 248, on July 14, 2026. The law requires AI companion operators to:
- disclose that the companion is not human
- build response protocols for suicidal ideation and self-harm
- file annual reports with the state Department of Health’s Behavioral Health Administration
4. Rhode Island Oversight of Artificial Intelligence Technology in Mental Health Care Act: Gov. Dan McKee signed the law, Senate Bill 2197 and House Bill 7349, on June 22, 2026. The law:
- requires informed consent from patients, or their parents or guardians, before AI companion models assist in therapy
- ensures licensed mental health professionals, not AI, make therapeutic decisions and create treatment plans
- applies existing confidentiality standards to AI tools used in mental healthcare, including notetaking tools
- bars companies and individuals from offering or advertising therapy unless a licensed professional provides it
5. Rhode Island chatbot safety law: Mr. McKee signed a second AI law, Senate Bill 2195 and House Bill 7350, on June 22, 2026. Sen. Lori Urso and Rep. Tina Spears sponsored both this law and the mental health law. The chatbot law requires operators to:
- build protocols for users who express suicidal ideation or self-harm and refer them to crisis services
- notify users that they are not talking to a human at the start of each session and at least every three hours
- file annual reports with the attorney general
6. Rhode Island AI scribe notification law: Mr. McKee signed a law June 22, 2026, requiring healthcare providers and facilities to notify patients when they use AI to document in-person or telehealth visits. Providers must also review the AI-generated documentation for accuracy. The legislation, House Bill 7538 and Senate Bill 2570, was sponsored by Sen. Pamela Lauria, MSN, and Rep. Teresa Tanzi.
7. Iowa House File 2635: Iowa Gov. Kim Reynolds signed the law on May 13, 2026, and it took effect July 1, 2026. Utilization review organizations may use AI for initial prior authorization reviews. However, AI cannot be the sole basis for denying, delaying or downgrading a request based on medical necessity. The law also prohibits health carriers from penalizing providers for out-of-network referrals.
8. Maine Legislative Document 2082: Gov. Janet Mills signed the law on April 13, 2026. It bars anyone from providing or advertising therapy to the public, including through AI, unless a licensed professional provides it. Licensed mental health professionals may use AI only for administrative and supplementary support. They may not let AI make therapeutic decisions, interact with clients in therapeutic communication, or generate treatment plans without their review.
When AI supports a session that is recorded or transcribed, the law requires providers to tell clients in writing that AI will be used, what it is for and how session data will be handled. Clients must also give consent.
9. Texas Responsible Artificial Intelligence Governance Act: Texas’ broader AI law took effect Jan. 1, 2026. Under the law, healthcare providers that use an AI system in a patient’s care must disclose that use to the patient or the patient’s personal representative. The disclosure is due no later than the date the service or treatment is first provided, or as soon as reasonably possible in an emergency. San Antonio-based University Health and Houston Methodist have built the disclosures into their existing consent processes.
10. California SB 243: Mr. Newsom signed Senate Bill 243 in October 2025, and it took effect Jan. 1, 2026. Under the law, which regulates companion AI chatbots:
- operators must disclose that users are interacting with AI
- minors get added protections, including recurring break reminders and limits on sexually explicit content
- operators must maintain protocols to prevent suicide and self-harm content and provide crisis referrals
- beginning July 1, 2027, operators must report annually to the Office of Suicide Prevention
11. California AB 489: Mr. Newsom signed Assembly Bill 489 in October 2025, and it took effect Jan. 1, 2026. The law bars AI tools from using titles, terms or phrases that falsely suggest a licensed human health professional is providing care or advice. Healthcare licensing boards enforce the law, and each improper use counts as a separate offense.
12. Texas Senate Bill 1188: Texas Gov. Greg Abbott signed the law in June 2025, and it took effect Sept. 1, 2025. Practitioners may use AI for diagnosis and treatment recommendations, but they must disclose its use to patients and review AI-generated records. The law also:
- requires covered entities to store patient EHRs in the U.S., with the storage requirement applying beginning Jan. 1, 2026
- requires parents and guardians to have access to minors’ records
- allows civil penalties of up to $5,000 per negligent violation, $25,000 per knowing violation and $250,000 when protected health information is used for financial gain
13. Illinois Wellness and Oversight for Psychological Resources Act: Gov. J.B. Pritzker signed the law on Aug. 1, 2025, and it took effect immediately. It prohibits AI tools from:
- making independent therapeutic decisions
- directly interacting with patients
- generating treatment plans without clinician review
- detecting emotions or mental states
Licensed professionals may use AI for administrative and supplementary support if they keep full responsibility and get informed patient consent. Violations carry fines of up to $10,000 each.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.