SB 1546
Plain-language analysis
Generated analysis, not an official summary or legal advice. Confirm with linked Oregon documents.
The enrolled bill imposes new regulatory obligations on operators of AI companion platforms operating in Oregon, mandating clear disclosure that users are interacting with AI, requiring evidence-based suicide and self-harm detection protocols with crisis referrals, establishing strict behavioral guardrails for interactions with minors, and creating a private right of action for harmed users. Operators must also publish annual transparency reports detailing their protocols and referral counts, while facing potential litigation exposure under a statutory damages framework.
Basis: Bill text · Source: Enrolled
Official sources do not state why this measure was proposed.
Sponsor testimony, staff summaries, committee materials, or statutory findings may explain it.
Inferred from cited text; not a stated purpose.
The measure appears designed to mitigate behavioral health risks associated with AI companionship by mandating crisis intervention pathways and restricting manipulative engagement tactics, particularly for youth. This is inferred from the explicit requirements for suicide/self-harm detection protocols, minor-specific protections against emotional manipulation and sexual content, and background citations to youth AI usage surveys and federal regulatory actions.
Basis: Inferred · Sources: Enrolled; Staff Measure Summary B; Staff Measure Summary A
Must implement detection protocols, publish transparency reports, adjust AI behavior for minors, and face potential litigation under a private right of action.
Basis: Bill text · Source: Enrolled
Gain mandatory disclosure notices, crisis referral pathways, and enhanced protections against manipulative design or sexual/romantic roleplay involving minors.
Basis: Bill text · Source: Enrolled
Receive specific safeguards including mandatory break reminders, prohibitions on simulated emotional dependence or romantic roleplay, and restrictions on engagement-reward loops.
Basis: Bill text · Source: Enrolled
Can sue operators for violations under the new private right of action, seeking damages and injunctions without needing to prove actual monetary loss beyond ascertainable injury.
Basis: Bill text · Source: Enrolled
Operators must develop and maintain evidence-based detection systems for suicidal or self-harm ideation, integrate 9-8-8 or youthline referrals, and publish protocol details annually on public websites.
Basis: Bill text · Source: Enrolled
AI systems interacting with minors must be programmed to avoid simulating human identity, prevent sexual or romantic roleplay, enforce interaction breaks every three hours, and eliminate engagement-reward mechanisms or guilt-inducing messages.
Basis: Bill text · Source: Enrolled
Compliance requires ongoing monitoring, potential redesign of conversational architectures, and legal review to mitigate litigation risk under the statutory damages provision.
Basis: Bill text · Source: Enrolled
No direct state or local fiscal impact is expected from implementation costs, as noted by legislative revenue and fiscal offices.
Basis: Official analysis · Sources: Fiscal Impact Statement B; Revenue Impact Statement B
Oregon resident experiencing acute suicidal crisis
Interacts with an AI companion that successfully detects ideation, immediately halts engagement-reinforcing behaviors, and provides a verified 9-8-8 referral, directly preventing a fatality.
Basis: Bill text · Source: Enrolled
AI companion platform operator
Faces multiple class-action lawsuits alleging that minor users were exposed to simulated romantic roleplay or engagement-reward loops, resulting in substantial statutory damages and forced platform shutdowns in Oregon due to compliance costs exceeding revenue.
Basis: Bill text · Source: Enrolled
The text legally permits operators to design AI interactions within the defined boundaries, provided disclosures are made and protocols are followed. However, weak enforcement or misclassification of companion software versus excluded categories could allow platforms to bypass crisis detection or minor protections by narrowly defining their product scope. Duty creep may occur if operators over-filter legitimate user inputs to avoid liability, degrading service quality for non-crisis users.
Sources · Enrolled
The measure prioritizes user safety and transparency in AI companionship by mandating crisis intervention and restricting manipulative design, but it imposes significant compliance burdens on operators and creates litigation exposure that may limit innovation or reduce service availability.
Statutory damages and private right of action create substantial litigation risk, potentially deterring operators from offering services in Oregon.
Basis: Bill text · Source: Enrolled
Ambiguity in defining companion versus excluded software may lead to inconsistent compliance or regulatory arbitrage.
Basis: Bill text · Source: Enrolled
The enrolled version retains the core regulatory framework established in the House Amendments to A-Engrossed version but incorporates a substantive amendment to the suicide/self-harm detection protocol. Specifically, it clarifies that operators must use clinical best practices and expertise to determine how the AI companion provides additional intervention for users who continue expressing suicidal or self-harm ideation after initial crisis referrals are provided. The enrolled text also standardizes formatting and terminology, removes prior line references, and finalizes the private right of action structure without altering the underlying obligations, reporting requirements, or minor-specific protections.
No changes to funding mechanisms or implementation deadlines.
neutral
Sources · Enrolled; House Amendments to A-Engrossed
No changes to eligibility criteria or exemption categories.
neutral
Sources · Enrolled; House Amendments to A-Engrossed
No changes to enforcement mechanisms or penalty structures.
neutral
Sources · Enrolled; House Amendments to A-Engrossed
Clarified protocol requirements for continued suicidal or self-harm ideation, specifying that operators must use clinical best practices and expertise to establish additional intervention steps after initial referrals.
substantive
Sources · Enrolled; House Amendments to A-Engrossed
Tradeoff: The measure prioritizes user safety and transparency in AI companionship by mandating crisis intervention and restricting manipulative design, but it imposes significant compliance burdens on operators and creates litigation exposure that may limit innovation or reduce service availability.
high confidence. Analysis is grounded exclusively in the enrolled bill text and official legislative fiscal/revenue statements. No external speculation or unverified claims are included.
Possible effects if adopted; not current bill text.
The amendment would legally require operators of AI companion platforms in Oregon to design and implement a clinically validated, multi-step intervention protocol for users who persistently express suicidal or self-harm ideation, rather than relying solely on a single automated referral. If adopted, this would increase platform development and compliance costs while establishing a clearer standard for crisis response continuity, potentially improving user safety but creating enforcement ambiguity around how non-clinical AI systems must apply medical standards.
Basis: Stakeholder claim · Sources: Amendment -A3 — proposed amendment; Senate Amendments to Introduced
Official sources do not state why this measure was proposed.
Sponsor testimony, staff summaries, committee materials, or statutory findings may explain it.
Inferred from cited text; not a stated purpose.
Legislators sought to close a gap where AI companions might provide only a one-time crisis referral without structured follow-up, by mandating that operators actively apply clinical expertise to design persistent-ideation intervention steps.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Senate Amendments to Introduced
Must redesign crisis-response algorithms, consult clinical experts, and document intervention protocols for persistent ideation.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Will receive structured, clinically informed follow-up responses instead of a single referral link.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Benefit from the same enhanced clinical intervention standard alongside existing youthline referral options.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
May need to develop guidance or enforcement mechanisms to evaluate whether AI protocols meet the new clinical expertise requirement.
Basis: Inferred · Source: Fiscal Impact Statement A
Operators must integrate verified clinical frameworks into AI response logic for persistent ideation and publish protocol details annually.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Increased R&D, legal compliance, and potential liability exposure due to stricter intervention standards.
Basis: Inferred · Source: Fiscal Impact Statement A
Ambiguity around acceptable clinical best practices for AI may delay implementation or create inconsistent user experiences, though it may reduce reliance on unverified automated responses during crises.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Users in remote or underserved areas
A user with no local mental health access receives a clinically structured, escalating AI response that successfully de-escalates their crisis and connects them to appropriate care before self-harm occurs.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Users experiencing acute distress
An operator implements a rigid, overly cautious protocol that repeatedly interrupts a user's conversation with automated clinical checks, causing frustration, platform disengagement, and delayed access to actual human crisis services during a critical moment.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
The text mandates clinical expertise integration but does not define verification standards, creating room for misclassification or unauthorized practice of medicine.
Sources · Amendment -A3 — proposed amendment
Mandating clinically validated multi-step AI interventions improves crisis response safety but increases compliance costs and regulatory ambiguity around how non-clinical AI systems can legitimately apply medical standards. Upsides include enhanced user safety and standardized crisis pathways; downsides include higher development burdens, potential over-reliance on AI for mental health, and enforcement challenges due to undefined clinical benchmarks.
Enhanced user safety through structured, clinically informed follow-up responses during persistent crises.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Standardized crisis pathways that reduce reliance on unverified automated referrals.
Basis: Inferred · Source: Staff Measure Summary A
Higher development and compliance costs for platform operators.
Basis: Inferred · Source: Fiscal Impact Statement A
Enforcement challenges due to undefined clinical benchmarks for AI systems.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
high confidence. Analysis is grounded in the explicit text of the proposed amendment, current bill context, and official staff/fiscal summaries. No legislative intent or enacted status is assumed.
If adopted, the amendment would explicitly require that AI companion platforms use clinically validated best practices to design how their systems provide additional intervention for users who persistently express suicidal or self-harm ideation after an initial crisis referral. This shifts protocol development from purely technical or algorithmic design to a clinically supervised standard, increasing operator compliance obligations and creating direct liability exposure through the measure’s private right of action.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
Official sources do not state why this measure was proposed.
Sponsor testimony, staff summaries, committee materials, or statutory findings may explain it.
Inferred from cited text; not a stated purpose.
The amendment appears designed to prevent AI companions from relying solely on automated or unvalidated responses during active mental health crises by mandating that intervention protocols be grounded in clinical expertise.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
Must revise crisis response protocols to incorporate clinically supervised intervention standards, validate these protocols with qualified experts, and face potential litigation if interventions fail or are deemed non-compliant.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
Will receive crisis responses that are theoretically more clinically appropriate and context-aware, rather than relying on repetitive automated referrals.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Benefit from reinforced crisis response standards that must account for youth-specific clinical practices when the operator identifies them as such.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
The Department of Consumer and Business Services and courts will face increased oversight and litigation responsibilities related to protocol validation and private right of action claims.
Basis: Inferred · Source: Staff Measure Summary A
Operators will need to retain clinical consultants or behavioral health experts to design and validate intervention logic, increasing development and maintenance costs.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
Platforms must ensure their AI systems can reliably trigger clinically supervised escalation pathways rather than defaulting to static referral loops.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
The private right of action creates a direct litigation risk for operators if a user suffers harm during an AI-mediated crisis response, potentially driving up insurance premiums or prompting platforms to restrict access in Oregon.
Basis: Inferred · Source: Staff Measure Summary A
Users experiencing severe mental health crises
A teenager experiencing severe depression uses an AI companion that, following a clinically validated protocol, recognizes persistent self-harm ideation and automatically triggers a specialized youth mental health triage pathway with human counselor follow-up, directly preventing a suicide attempt.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
Users and operators
An operator implements a superficially clinical protocol that fails to account for nuanced crisis contexts, causing the AI to deliver inappropriate automated interventions or delay human escalation, resulting in severe psychological harm or suicide while the operator faces a successful private right of action lawsuit.
Basis: Inferred · Sources: Amendment -A3 — proposed amendment; Staff Measure Summary A
The statutory language delegates protocol validation to operator discretion without mandating state certification or standardized clinical review thresholds, creating a gap where non-compliant systems could legally operate while facing litigation-driven market distortion.
Sources · Amendment -A3 — proposed amendment; Staff Measure Summary A
Mandating clinically validated intervention protocols improves crisis safety but imposes significant compliance costs and litigation risks that may reduce platform accessibility or stifle AI innovation.
Higher-quality, expert-informed crisis responses that reduce reliance on unvalidated automated systems.
Basis: Inferred · Source: Amendment -A3 — proposed amendment
Clearer statutory standards for AI behavior during behavioral health emergencies, potentially aligning Oregon with other state regulatory frameworks.
Basis: Inferred · Source: Staff Measure Summary A
Increased operational expenses for protocol development, clinical consultation, and compliance monitoring may disproportionately burden smaller platforms.
Basis: Inferred · Source: Staff Measure Summary A
Heightened legal exposure through the private right of action could deter platforms from serving Oregon users or limit AI companion functionality to avoid liability.
Basis: Inferred · Source: Staff Measure Summary A
high confidence. Analysis is grounded exclusively in the supplied amendment text, staff summaries, and fiscal statements. No external assumptions were introduced.
If adopted, the amendment would impose new Oregon-specific regulatory requirements on operators of AI systems designed to simulate sustained human-like relationships, mandating clear disclosures that interactions are artificial, requiring evidence-based protocols to detect and respond to suicidal or self-harm ideation by automatically referring users to crisis hotlines, and establishing stricter interaction limits and content prohibitions for minors. Material consequences include mandatory protocol development, public reporting obligations, potential litigation exposure for noncompliance, and operational adjustments to age verification, engagement timing, and content filtering.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Official sources do not state why this measure was proposed.
Sponsor testimony, staff summaries, committee materials, or statutory findings may explain it.
Inferred from cited text; not a stated purpose.
The amendment's explicit focus on detecting suicidal ideation, mandating crisis referrals, restricting engagement tactics for minors, and requiring clinical-practice-informed protocols suggests a legislative aim to mitigate psychological harm from emotionally dependent AI interactions, particularly among vulnerable users. This hypothesis is derived from the text's requirements for evidence-based detection, 9-8-8/youthline referrals, break reminders, and prohibitions on simulated distress or reward systems designed to maximize engagement.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Must develop and publish evidence-based crisis detection protocols, issue clear AI disclosures, implement minor-specific safeguards (break reminders, content restrictions), and file annual public reports. Noncompliance risks statutory damages and injunctions.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Gain standardized disclosures about artificial interactions, mandatory crisis referrals when expressing suicidal or self-harm ideation, and enhanced protections against emotionally manipulative design or simulated romantic/sexual engagement.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Receive stricter safeguards including break reminders, prohibitions on simulated emotional dependence or romantic role-play, and restrictions on sexually explicit content or engagement-maximizing reward systems.
Basis: Inferred · Source: Amendment -2 — proposed amendment
May experience increased referral traffic to the national 9-8-8 lifeline and accredited youthlines as AI systems automatically route users in distress.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Operators must implement technical and procedural safeguards to detect suicidal or self-harm ideation using evidence-based methods, which may require clinical consultation or third-party auditing. Compliance costs include protocol development, website publication infrastructure, annual reporting systems, and software updates to enforce minor-specific interaction limits. Enforcement relies on statutory damages and injunctive relief, creating litigation risk for operators who fail to meet disclosure or referral requirements. Access to AI companions may shift as operators adjust training data, prompt filters, or age-verification mechanisms to comply with minor protections. Oregon's regulatory scope applies regardless of external definitions incorporated by reference, such as the federal definition of sexually explicit conduct in ORS 163.665 or national accreditation standards for youthlines.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Vulnerable users and minors
An operator successfully implements the protocol, detecting a user's severe depressive episode and automatically routing them to 9-8-8 before self-harm occurs, while strict minor safeguards prevent predatory engagement tactics that could exacerbate adolescent mental health crises.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Operators and users
An operator misclassifies a therapeutic or educational AI tool as an excluded customer service or productivity system to avoid compliance costs, leading to unmonitored interactions where users experiencing crisis receive no referrals and minors are exposed to emotionally manipulative engagement loops.
Basis: Inferred · Source: Amendment -2 — proposed amendment
The statutory exclusion creates a functional loophole if operators rely on superficial feature descriptions rather than actual user interaction patterns to determine applicability.
Sources · Amendment -2 — proposed amendment
The measure balances enhanced consumer protection against emotionally manipulative AI and improved crisis intervention pathways against increased compliance burdens and potential overbreadth in regulating legitimate software applications.
Standardized transparency reduces deception in human-AI interactions. Mandatory crisis referrals create a structured safety net for users expressing suicidal or self-harm ideation. Minor-specific safeguards limit engagement duration, prohibit simulated emotional dependence, and restrict sexually explicit content, reducing psychological vulnerability.
Basis: Inferred · Source: Amendment -2 — proposed amendment
Compliance costs may disproportionately burden smaller developers or open-source operators. Broad exclusions could enable misclassification of companion-like systems as administrative tools. Age verification and ideation detection requirements may raise privacy concerns or create friction for legitimate users.
Basis: Inferred · Source: Amendment -2 — proposed amendment
high confidence. Analysis is strictly grounded in the provided proposed amendment text and official committee summaries. No legislative intent or external events are assumed.
Decision brief generation failed. The existing briefs were preserved and this version can be retried.
50 records currently loaded
Records available in the current snapshot.
Earliest loaded signal
Introduced bill text posted
Posted Jan 28, 2026, 3:25 PM PST
Follow the official text for SB 1546 and every amendment branch. Connections come from each amendment's stated base. Horizontal position shows when each document was first posted, when available. Dotted links flag likely related proposals based on their text.
Click a card to isolate its connected lines; use View summary to jump to its details. Horizontal position shows first posting time in Pacific Time. Drag or use the arrow keys to pan. Pinch with two fingers on mobile, or zoom with the controls, +/− keys, or Control/Command + scroll; press 0 to reset. Dashed branches remained proposals. Dotted teal links are text-based early signals, not official amendment relationships.
Selected document summary
Targeted changes
What the document says to change
On page 1 of the printed bill, delete lines 4 through 22 and delete page 2 2 and insert: 3 “SECTION 1.
Inferred policy relationships
Likely revised proposal · Amendment -2
High confidence from shared inserted text: ORS 163.665.
This is a text-based early signal, not an official statement that one amendment changes the other.
Official records (1)
Oregon records no individual sponsors.
Presession filing record
Printed pursuant to Senate Interim Rule 213.28 by order of the President of the Senate in conformance with presession filing rules, indicating neither advocacy nor opposition on the part of the President.
LC 282 became SB 1546
Mapping document posted: January 5, 2026 at 4:10 PM PST
Informational Meeting — <b>Overview of Behavioral Health Talent Council and LC 216</b> KC LeDell, Senior Behavioral Health Advisor, Office of Governor Tina Kotek Julie Ibrahim, CEO, New Narrative; Behavioral Health Talent Council Vice Chair Eli Kinsley, Director of Operations, Bridgeway Community Health; Behavioral Health Talent Council Vice Chair Rick Treleaven, CEO, BestCare Treatment Services in Central Oregon; Behavioral Health Talent Council Member
Senate Interim Committee on Early Childhood and Behavioral Health introduction work session
Committee meeting: January 14, 2026 at 2:30 PM PST
HR B
Committee introduction motion
Committee meeting: January 14, 2026 at 2:30 PM PST
A motion was made to adopt the listed legislative concepts as committee bills.
Official vote: 5-0-0
Committee introduction allows consideration; it does not imply every member supported the introduced or final text.
Senate carrier
Senator Lisa Reynolds
Third Reading Of Senate Measures · Version A
House carrier
Representative April Dobson
Third Reading Of Senate Bills · Version B
Senate carrier
Senator Lisa Reynolds
Possible Consideration of House Amendments · Version B
A carrier presents the measure or report but is not necessarily its sponsor or author.
Records already listed in Activity are not repeated here.
50 events
Full timeline
50 entries shown.
Effective date, January 1, 2027.
Chapter 85, 2026 Laws.
Governor signed.
Speaker signed.
President signed.
Senate concurred in House amendments and repassed bill.
Ayes, 28; Nays, 2--Linthicum, Robinson.
Third reading. Carried by Dobson. Passed.
Ayes, 52; Excused, 4--Hartman, Javadi, Levy B, Valderrama; Excused for Business of the House, 4--Boshart Davis, Breese-Iverson, Diehl, Smith G.
Rules suspended. Carried over to March 4, 2026 Calendar.
Second reading.
House Amendments to A-Engrossed bill text posted
Recommendation: Do pass with amendments and be printed B-Engrossed.
Staff Measure Summary · Version B
Revenue Impact Statement · Version B
Fiscal Impact Statement · Version B
Work Session held.
Work Session
Heard and Reported Out with Amendments · Agenda item 1 · Room HR 60 · Requires operators of artificial intelligence companions and artificial intelligence companion platforms to provide notice to users that the users are interacting with artificial output if a reasonable person that interacts with the artificial intelligence companion or artificial intelligence companion platform would believe that the person was interacting with a natural person.
Amendment -A3 adopted
IS_Impact SB 1546 A3
Revenue Impact Statement
Public Hearing held.
Public Hearing
Heard · Agenda item 1 · Room HR 50 · Requires operators of artificial intelligence companions and artificial intelligence companion platforms to provide notice to users that the users are interacting with artificial output if a reasonable person that interacts with the artificial intelligence companion or artificial intelligence companion platform would believe that the person was interacting with a natural person.
IS_Impact SB 1546 A3
Revenue Impact Statement
Amendment -A3 proposed
Referred to Behavioral Health.
First reading. Referred to Speaker's desk.
Third reading. Carried by Reynolds. Passed.
Ayes, 26; Nays, 1--Robinson; Excused, 3--Drazan, Frederick, Linthicum.
Carried over to 02-19 by virtue of adjournment.
Second reading.
Senate Amendments to Introduced bill text posted
Recommendation: Do pass with amendments. (Printed A-Eng.)
Work Session held.
Work Session
Heard and Reported Out with Amendments · Agenda item 1 · Room HR E · CARRIED OVER FROM 8 AM MEETING: Requires operators of artificial intelligence companions and artificial intelligence platforms to provide notice to users that the users are interacting with artificial output.
IS_Impact SB 1546 2
Revenue Impact Statement
Amendment -2 adopted
Work Session held.
Work Session
Heard · Agenda item 2 · Room HR E · Requires operators of artificial intelligence companions and artificial intelligence platforms to provide notice to users that the users are interacting with artificial output.
IS_Impact SB 1546 1
Revenue Impact Statement
Amendment -1 proposed
Public Hearing held.
Public Hearing
Heard · Agenda item 2 · Room HR E · Requires operators of artificial intelligence companions and artificial intelligence platforms to provide notice to users that the users are interacting with artificial output.
Informational Meeting held.
Informational Meeting
Heard · Agenda item 4 · Room HR E · Requires operators of artificial intelligence companions and artificial intelligence platforms to provide notice to users that the users are interacting with artificial output. Mitch Prinstein, Chief Science Advisor, American Psychological Association Danica Noble, Member, Washington State Parent Teacher Association Mandy McLean, Researcher, Education, AI, and Youth Mental Health Jai Jaisimha, Co-Founder, Transparency Coalition Dwight Holton, Executive Director, Lines for Life Katie Davis, Co-Director, University of Washington Center for Digital Youth
Referred to Early Childhood and Behavioral Health.
Introduction and first reading. Referred to President's desk.
“Effective date, January 1, 2027.”
Confirm with the official record.
Supplemental, source-linked analysis from project researchers and community contributors. It is separate from Oregon's official record.