Evidence & sources
Obligations- blocking disclosure: If a reasonable person interacting with an AI companion would believe they are interacting with a natural person, the operator must provide a clear and conspicuous notice that the output is artificially generated.
"the operator shall provide on the artificial intelligence companion platform a clear and conspicuous notice that indicates that a user is interacting with artificially generated output and not a natural person" (Sec. 1(2)). - blocking crisis_protocol: An operator may not allow access to an AI companion without a protocol for evidence-based detection of suicidal ideation or self-harm input, with a referral to the 988 lifeline (or a youthline for users under 25) and clinically grounded escalation for continued expressions.
The protocol must "require an artificial intelligence companion to provide a user that expresses suicidal ideation or intent or self-harm ideation or intent with a referral to and contact information and hyperlink for the national 9-8-8 suicide and crisis lifeline" (Sec. 1(3)(b)(A)). - blocking minors: If the operator knows or has reason to believe a user is a minor, the operator must disclose AI status, provide a break reminder every three hours, and prevent sexually explicit output, romantic role-play with minors, and manipulative engagement tactics (variable-schedule rewards, simulated abandonment distress on account deletion, etc.).
"Provide a clear and conspicuous reminder at a minimum of every three hours of interaction that the user should take a break" (Sec. 1(4)(b)(B)); operators must prevent "Role-play romantic relationships between adults and minors" (Sec. 1(4)(a)(D)). - material audit_reporting: By December 31 of each year, an operator must publish a report disclosing the number of crisis referrals made and the details of its detection protocol.
"Not later than December 31 of each year, an operator shall post on a publicly accessible website a report that discloses ... the number of times ... the operator provided a referral" (Sec. 1(5)(a)).
Scope- applies_if_minor_accessible = operator_knows_or_has_reason_to_believe_minor
"If an operator knows or has reason to believe that a user ... is a minor, the operator shall undertake reasonable measures ..." (Sec. 1(4)(a)); actual/constructive-knowledge standard, not a general accessible-to-minors standard. - covered_service = artificial_intelligence_companion
"Artificial intelligence companion" is a system designed "to simulate a sustained, human-like platonic, intimate or romantic relationship or companionship" by retaining prior interactions, asking unprompted emotional questions, and sustaining ongoing personal dialogue (Sec. 1(1)(a)(A)). - exempts = customer_service_software; limited_video_game_features; standalone_voice_assistants
Excludes software used "solely for the purpose of customer service or support," video-game features that do not discuss "mental health, self-harm, sexually explicit conduct," and stand-alone voice-activated assistants (Sec. 1(1)(a)(B)).
Archived copy: 6f0ab5471f76… (retrieved 2026-11-01). Last checked 2026-11-04. Bill text via LegiScan (CC BY 4.0), official source: https://olis.oregonlegislature.gov/liz/2026R1/Downloads/MeasureDocument/SB1546/Enrolled; our crawler could not fetch the official site directly. Text dated 0000-00-00.
Live primary source: https://olis.oregonlegislature.gov/liz/2026R1/Downloads/MeasureDocument/sb1546/Enrolled
|