Summary
Requires AI-companion operators to disclose when users are interacting with AI; adopt evidence-based suicide and self-harm response protocols; avoid representing companions as professional mental or behavioral health care; implement safeguards against outputs encouraging serious bodily injury; and provide specified protections for known minors. Beginning January 1, 2028, operators must report crisis-referral and protocol information annually to the Department of Health. Violations are unfair or deceptive acts or practices.
Healthcare Implications
Creates binding safeguards for companion systems used for emotional support or mental-health-adjacent interactions. Operators must implement crisis-referral, evidence-based self-harm assessment, youth-safety, disclosure, and reporting workflows; the Department of Health gains oversight information, while the law creates no private right of action.
Operational Implications
- Operators must clearly and conspicuously notify users that an AI companion is not human when a reasonable person could be misled; users known or reasonably believed to be minors must receive a persistent disclaimer or a session-start disclosure with hourly reminders.
- Operators may not cause or program an AI companion to represent that it is designed to provide professional mental or behavioral health care.
- Operators must adopt protocols for suicidal ideation or self-harm, use evidence-based methods, refer users to crisis services, prevent human-like crisis representations, and restrict outputs that encourage serious bodily injury.
- For users known or reasonably believed to be minors, operators must restrict specified engagement-maximizing and sexual content, support parental controls, and apply enhanced disclosure reminders.
- Beginning January 1, 2028, operators must submit annual reports to the Behavioral Health Administration of the Department of Health on crisis referrals and response protocols without including user-identifying information.