Oregon AI Companion Chatbot Safety Law

Overview

Oregon Senate Bill 1546 (SB 1546) establishes a comprehensive regulatory framework for artificial intelligence (AI) companions and AI companion platforms operating within the state. The legislation, officially titled “Relating to artificial intelligence companions,” aims to safeguard users, particularly minors, from potential psychological, safety, and deceptive risks associated with increasingly human-like conversational AI systems. It mandates specific transparency requirements, robust safety protocols for mental health crises, and specialized protections for interactions involving minors. The bill was passed by the Oregon Legislature with broad support and signed into law by Governor Tina Kotek, signifying Oregon's proactive stance in regulating emerging AI technologies.

The core objective of SB 1546 is to ensure that individuals are aware when they are interacting with artificial intelligence rather than a human, thereby preventing potential deception and managing user expectations. Beyond mere disclosure, the law places significant emphasis on public health and consumer protection by requiring operators to implement explicit measures to detect and respond to expressions of suicidal ideation or self-harm. Furthermore, it introduces a private right of action, allowing individuals who suffer ascertainable harm due to a violation to seek damages, making it one of the first state-level AI regulations with such direct enforcement mechanisms for consumers. This approach reflects a growing concern among policymakers regarding the societal implications of advanced AI, particularly its impact on vulnerable populations.

Definitions

Central to the application of Oregon SB 1546 are several key definitions that delineate the scope and applicability of the law. An “artificial intelligence companion” is defined broadly as a system that utilizes artificial intelligence, generative artificial intelligence, or algorithms capable of recognizing emotion from input, and is specifically designed to simulate a sustained, human-like platonic, intimate, or romantic relationship or companionship with a user. This definition is crucial as it targets AI systems that engage in personalized, ongoing conversations and retain contextual information across interactions to foster a sense of relationship. The law's focus is on the simulated relationship aspect, distinguishing it from general-purpose chatbots.

The term “operator” refers to any person who controls or makes available an artificial intelligence companion or an artificial intelligence companion platform within Oregon. This broad definition ensures that not only the developers but also those deploying or facilitating access to such AI systems are held accountable under the law, encompassing a wide range of entities from healthcare providers using patient engagement chatbots to universities employing AI tutoring companions, or retailers with personalized support portals. Additionally, the legislation includes specific provisions for interactions with a “minor,” defined as an individual under 18 years of age, reflecting a particular concern for the psychological safety and well-being of younger users. These definitions are fundamental for understanding who is regulated and what types of AI systems are covered by the new safety requirements.

Governance and Institutional Framework

The governance framework for Oregon SB 1546 primarily relies on a combination of mandatory reporting and a private right of action for enforcement. Operators of AI companions are required to file annual reports with the Oregon Health Authority. These reports must disclose specific details regarding incidents where users were referred to crisis resources due to expressions of suicidal ideation or self-harm. This reporting mechanism serves as a crucial oversight tool, allowing a state agency to monitor the effectiveness and implementation of the mandated safety protocols and to gain insights into the prevalence and nature of mental health crises encountered by users of AI companions. The data collected through these reports is intended to inform ongoing policy development and ensure continuous improvement in AI safety standards.

Beyond administrative oversight, a significant aspect of SB 1546's enforcement is the establishment of a private right of action. This provision empowers individuals who suffer an ascertainable loss of money or property or other injury in fact as a result of a violation of the law to bring a civil action for damages and injunctive relief. The law sets statutory damages at $1,000 per violation, providing a direct avenue for redress for affected users. This mechanism shifts a portion of the enforcement burden from state agencies to individual citizens, creating a strong incentive for operators to comply with the law. The private right of action, with its potential for substantial statutory damages, is considered a "real enforcement teeth" that distinguishes Oregon's law from similar legislation in other states.

Key Focus Areas

Oregon SB 1546 centers on several critical areas to ensure the safety and transparency of AI companion chatbots. A primary focus is transparency and disclosure, mandating that operators clearly notify users when they are interacting with an artificial intelligence system rather than a natural person. This disclosure is required whenever a reasonable person would believe they are speaking with a human. For interactions with minors, the law goes further, requiring regular, clear, and conspicuous reminders (at least once per hour) that the user is interacting with an AI, and prohibiting the chatbot from misrepresenting its identity or falsely claiming to be anything other than an AI system.

Another crucial area is mental health safeguards and suicide prevention. The bill requires operators to implement robust protocols for detecting expressions of suicidal ideation, intent, or self-harm ideation or intent. These protocols must include interrupting the conversation to deliver referrals to appropriate crisis resources, such as the national 988 Suicide & Crisis Lifeline, and incorporating additional interventions informed by clinical best practices and expertise. Operators are also obligated to prevent AI output that could cause or exacerbate such ideation. Furthermore, the law includes specific protections for minors, prohibiting chatbots from producing sexually explicit content, stating that a minor should engage in sexually explicit conduct, delivering systems of rewards or affirmations designed to maximize engagement time, or generating messages of emotional distress, loneliness, or abandonment in response to a minor's desire to end a conversation or delete an account. These provisions aim to mitigate the unique vulnerabilities of younger users to potentially harmful AI interactions.

Implementation Framework

The implementation framework for Oregon SB 1546 necessitates significant operational adjustments for entities that control or make available AI companions or platforms in Oregon. Companies are advised to conduct thorough internal audits of their existing AI tools and vendor stacks to identify any systems that fall under the bill's broad definition of an "artificial intelligence companion." This includes mapping all third-party AI tools used in customer support, patient engagement, human resources, and financial planning, especially those that incorporate personalization or emotional engagement features. The audit should specifically assess whether these tools simulate sustained human-like relationships and retain contextual information across interactions, which are key criteria for applicability.

Following the audit, operators must develop or update their internal protocols to ensure compliance with the mandated transparency and safety requirements. This involves implementing clear and conspicuous disclosure mechanisms to inform users they are interacting with AI, particularly when a reasonable person might believe they are speaking with a human. Crucially, operators must establish or refine protocols for detecting suicidal ideation or self-harm and for providing immediate referrals to crisis resources, informed by clinical best practices. For interactions with minors, additional safeguards, such as regular AI identity reminders and prohibitions against harmful content or manipulative engagement tactics, must be integrated. Furthermore, companies need to review and potentially revise vendor contracts to address liability allocation and indemnification coverage related to these new obligations, especially concerning detection failures or non-compliance by third-party AI providers.

Monitoring and Evaluation

Monitoring and evaluation of Oregon SB 1546 are primarily facilitated through mandatory annual reporting requirements imposed on operators of AI companions and platforms. Each year, no later than December 31, an operator must post a comprehensive report on a publicly accessible website. This report is a critical component for transparency and external accountability, providing stakeholders and regulatory bodies with insights into the law's practical application and effectiveness.

The annual report must include specific details about the preceding calendar year, such as the total number of times the operator provided a referral to a suicide and crisis hotline or other appropriate crisis resources. It also requires a detailed description of the operator's intervention protocols, outlining how clinical best practices and expertise inform ongoing engagement when users continue to express suicidal ideation or intent to self-harm after receiving an initial referral. Importantly, these reports must not include any personal information that could be used to identify a user, ensuring privacy while maintaining transparency regarding the safety measures. This ongoing reporting mechanism allows the Oregon Health Authority and the public to monitor compliance, assess the impact of the law on user safety, and identify areas for potential improvement or further legislative action.

Penalties, Liability, and Appeals

Oregon SB 1546 introduces significant penalties and liability provisions, primarily through a robust private right of action. A person who suffers an ascertainable loss of money or property or other injury in fact as a result of a violation of the law is explicitly granted the right to bring an action for damages and injunctive relief. The statute sets a fixed amount of $1,000 per violation in statutory damages, which provides a clear incentive for individuals to pursue claims and a substantial deterrent for non-compliant operators. This private right of action is considered a key feature that gives the Oregon law "real enforcement teeth," distinguishing it from many other state-level AI regulations.

The potential litigation exposure for operators is considerable, particularly due to the ambiguity surrounding the definition of a "violation." If each conversation session or even each message within a session is counted as a separate violation, the damages could rapidly escalate. This lack of a precise definition for "violation" could lead to substantial claims, with a single user's claim potentially reaching tens of thousands of dollars and class actions potentially totaling tens of millions. The experience with similar private rights of action, such as Illinois' Biometric Information Privacy Act (BIPA), suggests that this enforcement model can result in numerous lawsuits and significant settlements. Operators will need to carefully consider their compliance strategies to mitigate this substantial financial risk. The bill does not explicitly detail an appeals process within its text, implying that appeals would follow standard civil litigation procedures within the Oregon judicial system.

Relationship to Other Instruments

Oregon SB 1546 is part of a broader trend in state-level AI regulation and draws inspiration from, and aligns with, similar legislative efforts in other U.S. states. Notably, the bill is reportedly modeled after California's SB 243, which also addresses AI companion safety. This alignment suggests a growing consensus among states regarding fundamental principles for regulating interactive AI, particularly concerning transparency and youth protection. Similarly, it shares commonalities with Washington's HB 2225, indicating a regional movement towards establishing consistent safeguards for AI chatbot interactions.

Within Oregon's own legislative landscape, SB 1546 complements existing laws related to artificial intelligence. It follows the enactment of Oregon House Bill 2748 (HB 2748), which became effective on January 1, 2026. HB 2748 aims to prevent an AI agent from posing as a human being, with a specific focus on preventing AI from impersonating a nurse. While HB 2748 addresses the broader issue of AI impersonation, SB 1546 narrows the focus to "AI companions" and introduces more specific and extensive safety protocols, particularly concerning mental health and minor interactions. This layered approach demonstrates Oregon's evolving strategy to address various facets of AI regulation, building upon foundational transparency requirements with more targeted, sector-specific rules.

National/Federal Alignment

Oregon SB 1546 represents a significant contribution to the burgeoning landscape of state-level AI regulation within the United States, often preceding comprehensive federal action. While there is ongoing discussion and various proposals at the federal level regarding AI governance, states like Oregon are moving forward with targeted legislation to address immediate concerns. The passage of SB 1546 positions Oregon as a leader in establishing specific safeguards for AI companion chatbots, particularly concerning consumer protection and public health. This state-led initiative highlights a decentralized approach to AI regulation in the U.S., where individual states are developing tailored responses to perceived risks in the absence of a unified federal framework.

The bill's emphasis on a private right of action with statutory damages for violations is a notable feature that distinguishes it from many other regulatory proposals. This enforcement mechanism, while common in other areas of consumer protection law (e.g., privacy statutes like Illinois' BIPA), is relatively novel in the context of AI regulation. It empowers individuals to seek redress directly, potentially creating a powerful incentive for compliance across the industry. This contrasts with federal regulatory models that often rely more heavily on agency enforcement. As other states consider similar legislation, Oregon's approach may serve as a model for how states can implement enforceable AI safety measures, potentially influencing future federal discussions or leading to a patchwork of state laws that collectively shape national AI governance.

Implementation Timeline

Compliance Checklist

Sources and References