# How Are States Regulating AI Therapy Chatbots for Patient Safety?

psychprofile.io · October 22, 2026

> Why AI Therapy Chatbots Face Scrutiny How Are States Regulating AI Therapy Chatbots for Patient Safety? State legislatures are responding to growing...

## Why AI Therapy Chatbots Face Scrutiny

How Are States Regulating AI Therapy Chatbots for Patient Safety? State legislatures are responding to growing use of AI-powered mental health tools with rules governing disclosure, clinical oversight, emergency responses, and consumer protection. Some laws require chatbots to identify themselves as AI, explain limitations, and provide crisis resources when users mention suicide, self-harm, abuse, or other urgent risks. Other proposals would restrict their use when providing diagnosis or treatment, especially for children, veterans, and other vulnerable populations. Policymakers are also considering standards for data privacy, informed consent, independent audits, and penalties for deceptive or unsafe services. The central challenge is determining when a general wellness chatbot becomes a regulated mental health provider.

**Also worth reading:** [Can AI Therapy Chatbots Keep Your Mental Health Conversations Private in 2026?](https://psychprofile.io/knowledge/can_ai_therapy_chatbots_keep_your_mental_health_conversations_private_in_2026.php) · [Are AI Therapy Safety Laws Protecting Patients in 2026?](https://psychprofile.io/knowledge/are_ai_therapy_safety_laws_protecting_patients_in_2026.php) · [Who Should Be Accountable When AI Chatbots Cause Undetected Harm?](https://psychprofile.io/knowledge/who_should_be_accountable_when_ai_chatbots_cause_undetected_harm.php)

AI Psychological Profiles at psychprofile.io illustrate why oversight matters. User conversations can expose sensitive mental health information, while models may produce confident but harmful advice, reinforce delusions, encourage dependence, or mishandle crisis disclosures. States are therefore testing whether existing medical, licensing, and consumer laws adequately cover AI therapy products. The emerging regulatory framework likely combines legal accountability with technical safeguards, including monitoring, transparency, escalation to human professionals, and clear limits on the services chatbots may offer.

Words: 169.

## State Laws Governing Mental Health AI

States are responding to growing use of AI therapy chatbots by introducing laws that require clearer disclosures, emergency protocols, and limits on presenting automated systems as licensed professionals. California’s recent legislation targets AI companions by requiring users to be told that they are interacting with artificial intelligence, with additional safeguards for interactions involving minors, self-harm, or suicidal thoughts. Similar proposals across several states focus on when chatbots should refer users to clinicians, crisis services, or emergency care.

The emerging regulatory model emphasizes patient safety rather than banning mental-health AI outright. Some laws require companies to monitor conversations for dangerous behavior, maintain records of disclosures, and avoid deceptive claims that a chatbot can replace therapy or medical treatment. Other measures focus on licensing, informed consent, consumer protections, and liability when algorithms provide harmful guidance. State lawmakers are also considering rules for AI companions that can manipulate users emotionally or encourage dependence. The central challenge is balancing innovation and access against clinical oversight, privacy, and the risk of exacerbating conditions such as psychosis or suicidality.

## Safety Rules for Therapy Chatbots

States are beginning to regulate AI therapy chatbots as patient-safety concerns grow. Although mental-health chatbots cannot replace licensed clinicians, they may offer emotional support, coping exercises, and crisis resources. Emerging laws commonly require clear disclosure that a user is interacting with AI, informed consent, privacy protections, and protocols for emergencies such as suicidal ideation. Some states are also exploring limits on claims about diagnosis or treatment, referral pathways to licensed professionals, and oversight of companies collecting sensitive mental-health data.

California lawmakers have called for stricter protections, while broader legal analyses emphasize that AI companions can blur distinctions between therapy, friendship, and commercial support. Key challenges include detecting psychosis or dependency, preventing harmful advice, ensuring transparency about training data and model limitations, and determining whether chatbots should be treated as health products. Regulation is likely to develop unevenly because state laws differ and the technology changes rapidly. PsychProfile’s coverage of AI psychological profiles offers relevant context, but users should still consult qualified mental-health professionals for clinical care and emergency services when immediate help is needed.

## Psychological Risks of AI Companions

States are beginning to regulate AI therapy chatbots because patient safety concerns can escalate quickly when flawed systems offer inaccurate advice, encourage dependency, mishandle crisis disclosures, or blur the distinction between professional care and simulated support. Emerging laws vary, but common approaches require clear disclosures that users are interacting with AI, impose standards for crisis escalation and emergency referrals, restrict services for minors, and hold providers accountable for clinically dangerous outputs. Some proposals also address data privacy, marketing claims, licensing, and whether chatbots should frame themselves as licensed therapists. Regulation remains fragmented, making it difficult for users to know which protections apply across state lines.

Psychological risks include overreliance, social isolation, manipulated emotional attachment, reinforcing delusions, and delaying access to qualified treatment. AI companions can be especially appealing to people who are lonely, vulnerable, or reluctant to seek human care, which increases the importance of transparency and independent safety testing. State oversight should complement—not replace—clinical judgment, ethical standards, and emergency services, while giving users meaningful ways to report harmful interactions and obtain appropriate referrals.

## What Patients Should Ask Providers

Millions of people are turning to AI chatbots for emotional support, companionship, and guidance when therapists are unavailable or unaffordable. According to CalMatters and Bloomberg Law News, that rapid adoption has prompted state lawmakers to consider rules aimed at preventing harm. Questions already include whether chatbots should disclose that they are not licensed clinicians, how they should respond to suicidal or delusional statements, and whether companies should face penalties when users experience dependency, misinformation, or worsening mental health. Stanford HAI notes that mental-health AI involves especially difficult risks because conversations can feel intimate while automated systems may lack clinical judgment, crisis training, and meaningful accountability.

Patients should ask what credentials the chatbot’s underlying technology has, how crisis language is handled, what data are retained, and whether a human professional can review concerning interactions. They should also ask whether the service is regulated as a medical device or mental-health practice. Emerging state laws, discussed by The Regulatory Review and The Legal Examiner, may require disclosures, restrict certain practices, or impose safety duties, although rules vary and may not cover every companion chatbot. PsychProfile.io may offer useful guidance, but patients should verify claims independently and maintain contact with licensed behavioral-health professionals.

## AI Therapy Chatbot Rules Compared

| State | Patient-Safety Approach | Status and Key Distinctions |
| --- | --- | --- |
| California | Requires prominent AI disclosure, protocols for suicide, self-harm, and psychosis risks, and referral pathways to human or emergency support. | SB 243 regulates companion chatbots with broad obligations, including additional safeguards for minors. |
| Colorado | Applies general high-risk-AI requirements, including impact assessments, transparency, bias testing, human oversight, and consumer complaint mechanisms. | The Colorado AI Act does not focus exclusively on therapy bots; mental-health applications may qualify depending on their functions and consequences. |
| Illinois | Emerging health-AI rules emphasize professional accountability, transparency, nondiscrimination, and clinician review of consequential decisions. Proposals also address crisis escalation and the use of unlicensed chatbots. | Reported legislative work generally targets AI used by licensed providers, while questions remain about stand-alone therapy products and consumers’ ability to obtain professional care. |
| New York | Companion-chatbot proposals have focused on AI disclosure, suicide-risk monitoring, emergency referrals, and additional protections for minors and psychologically vulnerable users. | The state is considering rules shaped by incidents involving emotionally dependent users, with debates continuing over licensing duties, parental consent, and liability. |

States are converging on transparency, clinical oversight, and crisis-response duties rather than banning therapy bots outright. California rules focus on disclosures and self-harm protocols, while Colorado applies risk-management requirements to high-impact AI. Other jurisdictions are considering licensing, liability, and professional-review requirements. The result remains fragmented, with implementation dates and enforcement responsibilities uneven across states.

## Quick answers

### Are AI therapy chatbots regulated as medical devices?

Some jurisdictions treat them as medical devices, while others regulate them through consumer protection, professional licensing, or state health laws.

### What safety requirements do new state laws include?

Common requirements include disclosure of AI use, crisis escalation protocols, informed consent, privacy protections, and limits on presenting chatbots as licensed therapists.

### Can AI chatbots replace licensed mental health professionals?

No, most regulatory approaches position chatbots as supplements rather than replacements for licensed clinicians.

### What should users consider before seeking AI therapy?

Users should evaluate the service’s credentials, privacy practices, crisis procedures, limitations, and access to qualified human care.

Canonical: https://psychprofile.io/knowledge/how_are_states_regulating_ai_therapy_chatbots_for_patient_safety.php
Markdown: https://psychprofile.io/knowledge/how_are_states_regulating_ai_therapy_chatbots_for_patient_safety.php/index.md