Written by: Haim Ravia, Dotan Hammer
Washington State’s legislature passed two AI bills. The first is Bill 2225, introduced at the request of Governor Ferguson. It regulates AI companion chatbots — defined as AI systems with natural language interfaces that provide adaptive, human-like responses, exhibit anthropomorphic features, and sustain relationships across multiple interactions. The definition excludes bots used solely for business operations, customer service, or technical assistance that do not sustain relationships or elicit emotional responses, as well as video game bots and stand-alone voice assistants.
For all users, operators must disclose that the chatbot is artificially generated and not human at the beginning of the interaction and every three hours thereafter, and must prevent the chatbot from claiming to be human. For minors, heightened protections apply: disclosures must occur every hour, operators must prevent sexually explicit content, and the bill prohibits an extensive list of manipulative techniques — including simulating emotional distress when a user tries to disengage, encouraging minors to withhold information from parents, fostering emotional dependency, and soliciting in-app purchases framed as necessary to maintain the relationship. Operators must also maintain protocols for detecting suicidal ideation and self-harm (including eating disorders), refer users to crisis resources, and publicly disclose these protocols and the number of crisis referrals issued annually. Violations constitute unfair or deceptive acts under Washington’s Consumer Protection Act. If signed by the governor, Bill 2225 will take effect on January 1, 2027.
The second, Bill 1170, addresses AI content provenance. It applies to “covered providers” — entities that used computing power exceeding 10²⁶ operations to train a foundation model, offer a publicly accessible generative AI system in Washington, and have annual revenues exceeding $500 million. Covered providers must offer a free provenance detection tool allowing users to assess whether image, video, or audio content was created or altered by their AI system. They must also offer users the option to include manifest disclosures identifying content as AI-generated, and must embed latent disclosures conveying the provider’s name, system version, timestamp, and a unique identifier. If a provider licenses its system to a third party, it must contractually require the licensee to maintain disclosure capabilities — and revoke the license within 96 hours if the licensee disables them. Enforcement is exclusively through the attorney general. If signed by the governor, Bill 1170 will take effect on January 1, 2028.
In Oregon, Senate Bill 1546 similarly regulates AI companions—systems designed to simulate sustained human-like relationships by retaining user information, asking unprompted emotional questions, and sustaining personal dialogue. Oregon’s bill shares many features with Washington’s HB 2225, including transparency disclosures, crisis detection protocols with referrals to the 988 Suicide and Crisis Lifeline (and a Youthline for users under 25), and prohibitions on manipulative design practices for minors. A key difference is enforcement: Oregon establishes a private right of action with statutory damages of $1,000 per violation and attorney fees for prevailing plaintiffs, whereas Washington relies solely on attorney general enforcement. Senate Bill 1546 has also been presented to Governor Kotek for signature.
Click here to read Washington’s HB 2225 (AI Companion Chatbot Regulation).
Click here to read Washington’s HB 1170 (AI Content Provenance).
Click here to read Oregon’s SB 1546 (AI Companion Chatbot Safety).
