AI Frontier Institute · Story
Companion chatbots and the mental-health line
In one year, states went from restricting AI in therapy to a shared template for companion chatbots: disclose the machine, detect self-harm, refer to a crisis line, and protect minors.
August 1, 2025
Illinois keeps AI out of therapy
HB 1806, the Wellness and Oversight for Psychological Resources Act (Public Act 104-0054), bars anyone from offering therapy or psychotherapy, including through internet-based AI, unless a licensed professional delivers it. Licensed clinicians may use AI only for administrative and supplementary support.
October 13, 2025
California enacts SB 243 on companion chatbots
SB 243 required operators of companion chatbots to disclose that the user is talking to a machine when a reasonable person could be misled, to maintain protocols for suicidal ideation with referral to crisis services, and to add protections for minors. It took effect January 1, 2026. The disclose, detect, refer, and protect-minors pattern in the entries below follows the same shape.
March 31, 2026
Oregon signs SB 1546 on AI companions
Operators must notify users when they are interacting with artificial output, run a protocol that detects and prevents output that could cause suicidal or self-harm ideation with referral to a crisis lifeline, add disclosures for minors, and publish an annual report.
April 14, 2026
Nebraska adopts the Conversational Artificial Intelligence Safety Act
Sections 12 through 18 of LB 525 require disclosure to minor account holders, either as a persistent visible disclaimer or at the start of each session and at least every three hours, ban unpredictable engagement rewards for minors, bar sexually explicit output to minors, and require a crisis-referral protocol.
May 27, 2026
Connecticut folds companions into An Act Concerning Online Safety
Public Act 26-15 (SB 5) requires operators of artificial intelligence companions to detect expressions indicating a risk of suicide, self-harm, or imminent violence, prevent output that encourages them, refer users to resources including 988, never claim to be human, and apply stronger measures for users under eighteen.
June 22, 2026
Rhode Island regulates AI in mental-health treatment
S 2197 set a framework for how AI may be used in mental-health care, under the state's Behavioral Healthcare, Developmental Disabilities and Hospitals authority, closing the loop back to the clinical line Illinois drew first.
July 14, 2026
Hawaii enacts the Artificial Intelligence Disclosure and Safety Act
SB 3001 (Act 248) requires clear disclosure whenever a reasonable person might believe they are talking to a human, for minors either a persistent disclaimer or a reminder at session start and at least hourly, crisis-intervention protocols with referral to hotlines, and a ban on presenting a companion as a professional mental-health provider.