State artificial intelligence laws create a complicated picture for psychologists

Following media reports of serious harms associated with AI tools—including chatbots allegedly encouraging self-harm or suicide—and in the absence of a comprehensive federal law governing AI in health care, many state legislators have proposed bills intended to prevent such tragic outcomes.

But critics say many state bills on AI cast too wide a net, limiting the ability of licensed mental health providers to access evidence-based AI tools that have the potential to augment or increase access to care. Following the passage of such restrictive legislation in Illinois, many states are considering similar bills that could severely limit psychologists’ access to beneficial tools designed to enhance patient care.

“I think everyone agrees on the goal of keeping the public safe from high-profile harms associated with AI,” said Leanna Fortunato, PhD, a licensed clinical psychologist and APA’s director of digital health and innovation. “The nuance is how to do it in a way that targets these problems while still leaving the door open for high-quality clinical tools and research.”

While some of these state laws do take important steps toward protecting the public from potentially harmful consequences of AI, others make it unlawful for clinicians, researchers, and developers to use clinical tools such as AI-powered programs that flag patients at increased risk of self-harm, for example, or to prescribe patient-facing apps that provide mental health interventions for patients to use between therapy sessions. In some cases, these laws could prevent psychologists from using technologies that have been proven safe and effective in clinical trials—potentially including the next wave of FDA-cleared digital therapeutics (PDF, 207KB)opens in new window.

“The digital therapeutics approved so far largely don’t incorporate artificial intelligence, but all signs point to that coming soon,” Fortunato said. “These aren’t just things you can download from the app store, but tools that have been tested in clinical trials and have to be ordered by a qualified health care provider.”

As of July 30, 2026, 33 bills related to health care AI policy had been signed into law across 22 states. A number of states have passed laws regulating AI use by health care providers, while several states, including Illinois, Maine, Nevada, and Vermont, have enacted legislation that specifically regulates the use of AI tools by mental and behavioral health care providers. But the landscape is changing quickly. In the first half of 2026 alone, state legislatures introduced more than 280 bills related to AI and health careopens in new window.

For clinicians, it’s challenging to stay on top of developments related to AI and clinical practice, but it’s a topic clinicians in every state must stay informed about, said Rick Barnett, PsyD, a licensed psychologist in Stowe, Vermont who also chairs the legislative committee for the Vermont Psychological Association. Barnett provided testimony when his state was considering AI legislation in early 2026—testimony that helped lead legislators to narrow the scope of the law that was eventually passed.

“AI is here and integrated more and more into daily life and clinical practice,” Barnett said. “We need to keep on top of existing and new laws to protect our patients and our practices.”

Proposed regulations come at AI from many directions. Many of the laws have clear benefits, both for the public and for clinicians. For example, some legislation mandates that AI chatbots cannot misrepresent themselves as licensed mental health providers. Other regulations require developers to enact crisis protocols if a chatbot detects a person using language about self-harm, or to put protections in place to prevent children from being exposed to potentially harmful content.

Other encouraging laws focus on areas like patient data privacy or disclosures. In some proposed bills, AI-enabled tools would be required to disclose to users that they are not human, or to remind users that they can make mistakes.

But some mental health providers and other stakeholders say many of the proposed regulations go too far, particularly those that limit psychologists’ and other health care providers’ use of AI-enabled clinical tools.

That was the case in Vermont, where Barnett lobbied for changes in the state’s proposed AI legislation. “The legislation was so sweeping, it seemed like it went far beyond banning chatbots to saying that clinicians cannot use AI in any way, shape, or form, under any circumstances,” Barnett said. “Virtually every electronic health record software now has AI embedded in it, so are we not allowed to use EHRs (electronic health records) anymore?”

Another concern for Barnett was that without clear definitions of what a bill did and did not allow, legislation could preemptively ban valuable tools that haven’t been developed or deployed yet. In Vermont, the bill eventually passed but was amended to avoid such unintended consequences for psychologists who use AI tools responsibly, thanks to advocacy from the Vermont Psychological Association and others.

Still, the problem is repeating itself in state after state. Many proposed bills attempt to set conditions about how providers can use AI in their practices. Some, for instance, lay out conditions that many psychologists would applaud: specifying that a human psychologist must review any AI-generated treatment materials, for instance, or informing patients about the use of AI tools and getting informed consent. But other proposed laws take a more restrictive approach, stipulating that AI tools can only be used for administrative tasks, while restricting “therapeutic communication”—essentially, any AI tools that interact directly with patients. That could severely limit the ability of psychologists to use beneficial tools designed to enhance patient care.

Many of these more restrictive pieces of legislation are modeled after a law passed in Illinois in 2025. That law prohibits AI tools from making independent therapeutic decisions, directly engaging in therapeutic communication with patients, and detecting patients’ emotions or mental states, even with a patient’s consent. Other uses, such as AI safety tools that alert a therapist to patients using phrases in session that may signal a risk of self-harm, fall into a legislative gray area under the law, according to the Artificial Intelligence in Mental Health Services Under Illinois Public Act 104‐0054: Legal Boundaries and a Framework for Establishing Safe, Effective AI Toolsopens in new window report in JMIR Mental Health.

“The Illinois law was notable in that it was one of the first to focus specifically on mental health professionals and how they can and can’t use AI tools,” Fortunato said. “Now we’re seeing a host of [proposed] laws modeled after the Illinois legislation.”

Critics of such bills say their overly broad language limit psychologists’ ability to use helpful tech tools—potentially including FDA-cleared digital therapeutics.

Legal hurdles are already impacting the development of such tools, said Nicholas Jacobson, PhD, an associate professor of biomedical data science and psychiatry at Dartmouth College who studies technology to enhance mental health assessment and treatment. He’s studying AI tools that collect passive behavioral data from smartphone sensors to predict depression recurrence and also developing systems that use generative AI to help patients access evidence-based psychotherapy interventions when they cannot connect with human psychologists.

He’s found evidence that such tools can provide safe and effective interventions for depression, anxiety, and eating disorders. But the legal landscape has been an obstacle to moving the work forward. When Illinois passed its AI bill, Jacobson was just getting ready to launch a large multisite trial of a generative AI tool for mental health. His collaborators were located in Illinois, and the team planned to recruit patients in several hospitals there. “We had to shut down the trial, because we did not believe we could proceed with the research in Illinois under that law,” Jacobson said.

There are other points of contention from critics of AI regulation. With differences from state to state, there are open questions about how state licensing boards will regulate and manage the fallout from the laws. And the bills don’t always reflect a clear understanding of how psychologists practice, said Deborah Baker, JD, head of APA’s Office of Professional Legal and Regulatory Affairs.

Some bills, for instance, aim to limit the use of AI for health care purposes—yet carve out exceptions for physicians. “That sends a message that the technology is so dangerous that mental health providers should only be able to use it in very limited ways, while it’s okay for physicians to use it,” Baker said. Practically, such carve-outs would create challenges for psychologists working in integrated systems such as hospitals or integrated care settings, who may not be able to access the same tools their physician colleagues are using. But another concern, Baker added, is the message that lack of parity sends about the role of clinical psychology in the health care space.

As the legal landscape evolves, APA is collaborating with state partners to stay abreast of newly proposed legislation and to continue advocacy efforts. A key goal, Fortunato said, is helping to ensure that proposed legislation is written with input from people with real-world knowledge about how psychologists work, who can reframe the focus to include meaningful safeguards and oversight of the technology.

APA staff have also testified to Congress about AI safeguards and met with the U.S. Food and Drug Administration officials as they consider how to regulate generative AI tools.

APA is also coordinating with other organizations such as the American Telemedicine Association to monitor what’s happening at the state level and align on messaging that will support practitioners’ responsible use of these tools.

Keeping track of all the potential changes may be even more challenging for psychologists involved with the Psychology Interjurisdictional Compact, since they are governed by different laws in each state where they treat patients.

To stay current on the latest laws and recommendations, Fortunato recommends psychologists connect with their state psychological associations. If you have questions about the implications of proposed legislation in your state, email APA’s Office of Professional Legal and Regulatory Affairs.