[MHAIP]

Building & policy for AI in mental health.

States are passing AI mental health laws faster than the evidence, and often without input from clinicians, consumers, or developers. We close the gap through research, analysis, stakeholder consultation, and model legislation that protects consumers and helps policies work in practice.

Subscribe for updates
50-state tracker · Live
20 enacted 17 pending 41 tracked
Enacted Pending Did not pass
Explore the tracker →

What's at stake

The use of AI for mental health support is already widespread.

The evidence on its impact is growing, showing both real benefits and real harms. The current patchwork of state laws, often written before that evidence was available, is not yet equipped to address either side of the picture.

0%

of US adolescents and young adults now use generative AI for mental health advice.

0+

AI-related bills introduced across 45 states by March 2026, with mental health a consistent focus.

0

enacted state laws governing AI in mental health, often using incompatible approaches.

A growing body of evidence documents AI-associated harms in users seeking mental health support, including emerging clinical reports of AI-related psychiatric presentations and crisis-detection failures linked to teen suicides. At the same time, AI tools have shown efficacy for mild-to-moderate mental health symptoms and have the potential to address real treatment gaps and reduce administrative burden on a strained mental health workforce.

Sources: McBain et al., 2025 · Østergaard, 2023 · Garcia v. Character Technologies (Settled, 2026) · Heinz et al., 2025 · Li et al., 2023 · Tierney et al., 2024

THE ISSUE

A new field of policy, still finding its footing.

States are writing the laws that will govern AI in mental health for at least the next decade. The work is just beginning, and three patterns are starting to define the landscape.

Challenge 1

Inconsistency

State approaches vary widely, and the inconsistencies are starting to matter.

States that have enacted AI mental health laws have taken different paths: bans on AI-led care and on clinician use of AI (Illinois, Maine), regulation of the companies building these tools (Utah), and frameworks combining both (Nevada). Each is a reasonable starting point. But the resulting variation now produces real cross-state effects.

Key findingThe same clinical use of AI is prohibited in some states and permitted in others.

Challenge 2

Evidence gap

The evidence base is still catching up to the laws being written.

States are writing AI mental health laws on legislative timelines while the evidence on these tools is still emerging. Some laws encourage safety protocols without specifying what they should contain. Where laws mandate crisis detection, they often do not require performance indicators, clinical oversight, or third-party testing.

Key findingMost chatbot systems struggle with crisis detection in empirical testing.

Challenge 3

Federal uncertainty

Federal developments are adding new questions to an evolving picture.

A December 2025 federal executive order has raised questions about how state protections interact with federal direction, even as a coherent federal framework for AI in mental health is still taking shape. The legislative cycle ahead will be formative for how this technology is governed nationwide.

Key findingDecisions made now will shape the policy environment for at least the next decade.

Our work

Four aims. Concrete deliverables.

The Mental Health AI Policy Project is an independent research and policy initiative working to ensure that state regulation of AI in mental health is grounded in evidence and shaped by the people it affects, consumers, clinicians, and developers, alongside the legislators writing the laws. We identify what every state is doing, analyze the real-world consequences for the people these laws affect, convene clinicians, consumers, and developers whose expertise is missing from current policymaking, and translate what we learn into model legislation and frameworks that states can adopt.

Identify

Map the current regulatory landscape.

We identify and track every state law and bill governing AI in mental health, creating a single source for a rapidly evolving regulatory landscape.

DeliverableA live 50-state tracker showing what each state requires of consumers, clinicians, and developers.

Analyze

Surface the real-world impact.

We compare state approaches with federal frameworks, empirical evidence, and international precedent.

DeliverableAnalytical briefs and comparative analyses that inform regulatory choices.

Convene

Bring affected voices to the table.

We convene clinicians, developers, and consumers—especially children and youth, individuals with disabilities, and historically marginalized communities—in structured consultations to surface lived experiences and frontline insights that are too often absent from state policymaking.

DeliverablePublished consultation summaries that inform our analyses and model legislation.

Translate

Translate evidence into policy.

We translate evidence into policy by developing model legislation and best-practice frameworks states can adopt, strengthening safety, clarifying clinician expectations, and enabling responsible innovation.

DeliverableModel legislation and best-practice frameworks states can adopt.

50-state tracker

A live map of state AI mental health legislation.

Our 50-state tracker compiles every active bill and enacted law governing AI in mental health, with details on scope, requirements, enforcement, and date of action.

Open the tracker

Our Commitments

The Mental Health AI Policy Project is independently governed. We publicly disclose our funders and publish our methodology and analyses.

Stay informed.

Monthly updates on state legislation, federal policy, and emerging research at the intersection of AI and mental health.

Contact us

We welcome conversations with funders, state policymakers, clinical organizations, and researchers working at this intersection.

info@mhaipolicy.org