The bill creates a new chapter in Title 19 RCW and adds a new section to chapter 4.20 RCW. It imposes duties on operators of “companion chatbots” and companion chatbot platforms: notify users when a chatbot could be mistaken for a human; maintain and publish protocols to prevent production of suicidal ideation, suicide, or self-harm content and to provide crisis service referrals (including referrals to 988 or its successor) when users express such thoughts; use evidence-based methods for measuring suicidal ideation; and make annual reports to the Washington State Department of Health beginning July 1, 2027, on crisis referrals and suicide-response protocols without including personal identifiers. For users the operator knows are minors, the bill requires disclosure that the interaction is with AI, default reminders at least every three hours during continuing interactions, and measures to prevent sexually explicit visual material or exhortations toward sexually explicit conduct. The bill also allows civil actions for persons who suffer an injury in fact from violations, authorizes injunctive relief, reasonable attorneys’ fees and costs, and damages equal to the greater of actual damages or $1,000 per violation. Legally, the change is a new regulatory regime plus changes to civil liability and wrongful-death procedure: it creates a new statutory chapter with operator obligations and reporting requirements, adds a prima facie evidence standard in chapter 4.20 RCW for suicides where an AI system received inputs and provided instructions/encouragement or failed to refer to 988 and the owner knew or should have known of the risk, and prohibits defendants from asserting that the AI autonomously caused the harm as a defense in related civil actions. The bill defines “artificial intelligence” and “companion chatbot,” and exempts certain bots (customer service/business bots, specified video game features, and certain voice assistants). Affected parties include the Department of Health, operators of companion chatbot platforms, owners/developers/modifiers/users of AI systems, users including minors, and crisis service providers. The extracted text omits specific RCW section numbers, does not specify how an operator must determine or verify that a user “is a minor,” and does not define the detailed standards for “reasonable measures,” “evidence-based methods,” or required protocol elements beyond referrals and detection/removal/response measures. The text also does not describe enforcement mechanisms beyond the civil causes of action and stated damages.

If enacted, companies that run or host “companion” chatbots will have to invest in written, public safety protocols, detection and removal systems for self-harm content, frequent on-screen reminders and bans on sexually explicit material for users they know are minors, and annual de-identified reporting to the Department of Health starting July 1, 2027. Those operators face new financial and legal risk because people harmed by violations can sue for injunctive relief and at least $1,000 per violation (or actual damages if higher), developers and owners can’t simply blame the AI as a defense, and a wrongful-death prima facie rule could make liability easier to prove in suicide cases where the system gave harmful instructions or failed to refer to 988. The Department of Health will take on the administrative work of receiving reports and publishing aggregate data, and crisis lines may see more referrals from chatbots. Practical questions that remain and will affect costs and operations include how operators are supposed to determine a user is a minor, what counts as “reasonable measures” or “evidence-based” methods, and the specific elements required in posted protocols, so companies will likely need legal and technical work to decide how to comply.