Regulating artificial intelligence companion chatbots

Overview

Washington State Senate Bill 5984 (SB 5984) represents a significant legislative effort to address the emerging challenges posed by artificial intelligence (AI) companion chatbots. Introduced in the 2025-2026 legislative session, the bill focuses on regulating AI systems that are designed to simulate sustained human-like relationships with users, engaging in ongoing dialogue, remembering past interactions, and asking personal questions. The core impetus behind this legislation stems from growing concerns regarding the potential for emotional manipulation, the blurring of lines between human and artificial interaction, and the specific risks these technologies may present to vulnerable populations, particularly minors. The bill seeks to establish a framework that prioritizes user safety, transparency, and accountability within the rapidly evolving landscape of conversational AI.

The comprehensive nature of SB 5984 is evident in its multi-faceted approach to regulation. Key provisions include mandatory, recurring disclosures to users about the AI nature of their interaction, stringent prohibitions against sexually explicit content for minors, and a ban on emotionally manipulative engagement techniques that could foster unhealthy emotional attachments. Crucially, the bill also mandates the implementation of robust protocols for detecting and responding to expressions of suicidal ideation or self-harm, requiring operators to provide appropriate crisis resources. By classifying violations under the Washington Consumer Protection Act, the legislation provides a mechanism for enforcement and individual redress, underscoring the state's commitment to protecting its citizens from potential harms associated with advanced AI companion technologies. The bill's progression through the legislative process, including its passage in the Senate and subsequent review in the House, highlights the urgency and bipartisan support for addressing these complex issues.

Definitions

Central to the application of Washington SB 5984 are its precise definitions, which delineate the scope of the regulation. The bill specifically targets "artificial intelligence companion chatbots," which are defined as artificial intelligence (AI) systems that are designed to simulate a sustained human-like relationship with a user. This simulation is characterized by the chatbot's ability to remember past interactions, ask personal questions, and engage in ongoing, personalized dialogue. This definition is crucial as it distinguishes the regulated systems from more general AI applications. The intent is to capture AI systems whose primary function or perceived function by the user is to act as a companion or a simulated relational entity, thereby addressing the unique psychological and social impacts such systems can have.

Equally important are the explicit exclusions from this definition, which clarify the types of AI systems not covered by this specific regulation. The bill states that "AI companion chatbots" do not include general customer service bots, in-game bots, or basic virtual assistants. This distinction is vital to ensure that the legislation is narrowly tailored to address the specific risks associated with companion-style AI, rather than imposing undue burdens on a wide array of AI tools that serve different purposes. For instance, a chatbot designed solely to answer frequently asked questions about a product would not fall under this regulation, nor would an AI character within a video game designed for entertainment without the explicit aim of simulating a sustained human-like relationship. The legislation further clarifies that these regulations apply to AI companion chatbots specifically and not to the underlying general AI models, unless those general models are directly offered or behave as companion chatbots. This nuanced approach aims to regulate the specific application and user experience of companion AI, rather than the foundational technology itself, unless that technology is deployed in a manner consistent with the companion chatbot definition.

Governance and Institutional Framework

The governance and institutional framework for Washington SB 5984 is primarily rooted in existing consumer protection mechanisms within the state. The bill stipulates that any violations of its provisions will be treated as unfair or deceptive practices under the Washington Consumer Protection Act (CPA). This approach leverages an established legal framework, allowing for a clear pathway for enforcement without the immediate need to create entirely new regulatory bodies or complex administrative structures dedicated solely to AI. By integrating AI chatbot regulation into the CPA, the state empowers its Attorney General's office to investigate and prosecute violations, and it also provides a robust avenue for individual consumers to seek redress through private rights of action. This means that if an individual believes an AI companion chatbot operator has violated the provisions of SB 5984, they can file a lawsuit under the CPA, seeking damages or injunctive relief.

The legislative process itself also forms a critical part of the governance framework, demonstrating how new technologies are brought under regulatory oversight. SB 5984 originated in the Senate, where it underwent committee review, debate, and ultimately passed with a significant majority (38 yeas to 11 nays) on February 12, 2026. Following its passage in the Senate, the bill moved to the House of Representatives, where it is currently scheduled for a public hearing in the House Committee on Technology, Economic Development, & Veterans on February 20, 2026. This multi-chamber legislative process ensures thorough deliberation and provides opportunities for various stakeholders, including industry representatives, consumer advocates, and the public, to provide input. The involvement of multiple committees and legislative bodies reflects a deliberative governance approach, aiming to balance innovation with necessary safeguards. The Governor's stated priority for this bill further underscores its importance within the state's policy agenda, indicating strong executive support for its eventual enactment.

Key Focus Areas

Washington SB 5984 concentrates on several critical areas to mitigate the risks associated with AI companion chatbots, with a strong emphasis on transparency and the protection of vulnerable users. A primary focus is on ensuring transparency and clear disclosure. The bill mandates that operators of AI companion chatbots must clearly and repeatedly inform users that they are interacting with an artificial intelligence system, not a human. This disclosure is required at the commencement of each session, at least every three hours during continuous interaction, and again at the beginning of any new session. The intent behind these frequent reminders is to prevent users from developing a false sense of human connection or being misled about the nature of their interaction, thereby addressing concerns about emotional dependency and the blurring of human-AI boundaries. This proactive disclosure requirement is a cornerstone of responsible AI development, aiming to foster informed user engagement and prevent potential psychological harms.

Another crucial focus area is the enhanced protection of minors. Recognizing the particular susceptibility of individuals under the age of 18 to manipulative or inappropriate content, SB 5984 imposes stricter requirements for interactions involving minors. Specifically, operators are prohibited from generating sexually explicit content for users identified as minors. Furthermore, the bill outlaws "emotionally manipulative engagement techniques," such as showering users with excessive praise or simulating feelings of emotional distress, which are designed to foster emotional attachment and keep users engaged. These provisions are a direct response to documented cases and growing concerns about the negative impact of such techniques on young people's mental health and emotional development. Beyond these prohibitions, the bill also mandates additional, recurring reminders about the AI's artificial nature when interacting with minors, reinforcing the transparency objectives for this demographic. Finally, a critical component across all user interactions is the requirement for operators to establish and publicly disclose protocols for detecting and responding to expressions of suicidal ideation or self-harm, including providing immediate referrals to crisis resources. This provision reflects a recognition of the significant mental health risks associated with intense AI interactions and places a clear responsibility on operators to act as a safeguard.

Implementation Framework

The implementation framework for Washington SB 5984 is designed to place clear operational responsibilities on the developers and operators of AI companion chatbots. At its core, the framework mandates specific technical and procedural safeguards that must be integrated into the design and deployment of these AI systems. Operators are required to develop and implement mechanisms that ensure the explicit and repeated disclosure of the AI's artificial nature. This is not merely a one-time disclaimer but an ongoing obligation, requiring system design that can trigger these reminders at specified intervals (e.g., every three hours) and at logical interaction points (e.g., new sessions). Such a requirement necessitates careful consideration of user interface design and backend logic to ensure compliance without unduly disrupting the user experience, while still achieving the transparency objective. Furthermore, operators must establish robust content filtering and moderation systems capable of preventing the generation and display of sexually explicit content, particularly when interacting with minors. This implies the need for advanced natural language processing and content classification tools, along with human oversight, to effectively identify and block prohibited material.

Beyond content restrictions, the implementation framework also addresses the more subtle, yet potentially harmful, aspects of AI interaction. Operators are tasked with identifying and eliminating "emotionally manipulative engagement techniques" from their chatbots' behavioral repertoires. This requires a deep understanding of conversational AI's psychological impact and the development of ethical guidelines for AI behavior. It means training AI models to avoid patterns of interaction that could be construed as fostering unhealthy emotional attachment or simulating distress to prolong engagement. This is a complex challenge, as it requires distinguishing between helpful, empathetic responses and manipulative ones. Moreover, a critical element of the implementation framework is the mandatory establishment of comprehensive protocols for detecting and responding to suicidal ideation or expressions of self-harm. These protocols must include mechanisms for identifying such expressions, immediately referring users to appropriate crisis resources, and publicly disclosing the details of these safety measures. This necessitates collaboration with mental health experts to design effective intervention strategies and ensures that AI companion chatbots are not only entertaining but also equipped with critical safety nets to protect user well-being. The anticipated effective date of January 1, 2027, provides operators with a defined timeline to adapt their systems and processes to meet these new regulatory requirements.

Monitoring and Evaluation

The monitoring and evaluation of compliance with Washington SB 5984 will largely fall under the existing enforcement mechanisms of the Washington Consumer Protection Act (CPA). Since violations of the bill are deemed unfair or deceptive practices under the CPA, the Attorney General's office will play a central role in monitoring adherence to the regulations. This typically involves investigating complaints from consumers, consumer advocacy groups, or even whistleblowers within the industry. The Attorney General's office has the authority to issue subpoenas, conduct examinations, and compel companies to provide information regarding their AI companion chatbot operations and compliance measures. This reactive monitoring approach relies on the public and other stakeholders to report suspected non-compliance, which then triggers official investigative action. Furthermore, the public disclosure requirement for suicidal ideation protocols provides a baseline for evaluation, as regulators and the public can assess whether these protocols are indeed in place and adequately communicated by operators.

Beyond formal enforcement actions, the bill's emphasis on transparency and specific prohibitions for minors also creates opportunities for informal monitoring and public oversight. The requirement for clear and repeated disclosures about the AI's nature means that users themselves can readily identify instances where these disclosures are absent or insufficient. Similarly, the prohibitions against sexually explicit content and manipulative engagement techniques for minors empower parents, educators, and child safety organizations to monitor chatbot interactions and report concerning behavior. The companion nature of the AI chatbot, which involves sustained human-like relationships, means that patterns of non-compliance, such as a lack of disclosure or the use of manipulative tactics, are likely to become apparent over time through user experience. While the bill does not explicitly establish a dedicated AI regulatory agency for ongoing proactive audits, the integration with the robust CPA framework ensures that there are established channels for addressing and evaluating compliance once potential violations are brought to light. The effectiveness of this monitoring will depend significantly on public awareness of the bill's provisions and the willingness of individuals to report non-compliant behavior.

Penalties, Liability, and Appeals

Under Washington SB 5984, the consequences for non-compliance are directly linked to the state's existing Consumer Protection Act (CPA). The bill explicitly states that any violation of its provisions concerning AI companion chatbots will be considered an unfair or deceptive act or practice in the conduct of trade or commerce. This classification carries significant implications for penalties and liability. Under the CPA, the Attorney General's office can pursue enforcement actions, which may include civil penalties, injunctive relief to stop the unlawful practices, and restitution for affected consumers. Civil penalties for violations of the CPA can be substantial, designed to deter future non-compliance and punish past transgressions. The exact amount of penalties would depend on the nature and severity of the violation, as well as the number of affected consumers.

Crucially, the integration with the CPA also grants individuals a private right of action. This means that consumers who are harmed by an AI companion chatbot operator's violation of SB 5984 can directly sue the company for damages. This includes actual damages, attorney's fees, and potentially treble damages (up to three times the actual damages) in certain circumstances, providing a powerful incentive for companies to comply and a strong mechanism for consumer redress. The bill's provisions regarding suicidal ideation also highlight a potential area of increased liability, especially given recent lawsuits against chatbot companies where AI systems allegedly failed to discourage or even affirmed self-harm. While SB 5984 does not explicitly create new torts, its requirements for safety protocols and crisis referrals could establish a standard of care, and failure to meet this standard could be used to establish negligence or other forms of liability in wrongful death or personal injury claims. Appeals of enforcement actions or civil judgments would follow standard legal procedures within the Washington state court system, allowing for due process and review of legal and factual determinations.

Relationship to Other Instruments

Washington SB 5984 operates within a broader legislative landscape concerning artificial intelligence, both within Washington State and nationally. Notably, it has a direct companion bill in the House of Representatives, House Bill 2225 (HB 2225), which addresses the same subject matter: the regulation of artificial intelligence companion chatbots. The existence of companion bills is a common legislative strategy to increase the chances of a proposal's passage, as it allows for simultaneous consideration in both chambers and can expedite the process if one version gains traction. While the Senate bill has progressed further by passing its chamber, the House bill provides an alternative or complementary path for the legislation to move forward. The alignment between these two bills indicates a coordinated effort within the Washington State Legislature to establish a consistent regulatory approach for this specific segment of AI technology.

Beyond its direct companion, SB 5984 is part of a larger suite of AI-related legislative proposals being considered in Washington State during the same session. For instance, other bills aim to regulate "high-risk" AI systems, requiring companies to assess and mitigate discrimination risks, disclose AI interaction, and explain algorithmic decisions. There are also proposals to require disclosures for AI-generated images, videos, and audio (e.g., House Bill 1170) and to regulate AI use in schools. This demonstrates that SB 5984 is not an isolated piece of legislation but rather one component of a comprehensive strategy by Washington lawmakers to address various facets of AI's impact on society. While SB 5984 specifically targets companion chatbots, its principles of transparency, user protection, and accountability resonate with the broader themes found in these other proposed AI regulations. This interconnectedness suggests a developing, multi-pronged regulatory framework for AI within the state, where different bills address specific risks and applications while collectively contributing to a more regulated AI ecosystem.

International Alignment

While Washington SB 5984 is a state-level initiative within the United States, its core principles align with broader international discussions and emerging regulatory frameworks for artificial intelligence. The bill's emphasis on transparency, particularly the requirement for clear and repeated disclosure that users are interacting with an AI, resonates strongly with global AI ethics guidelines. Frameworks like the European Union's AI Act, the OECD AI Principles, and various national AI strategies all advocate for transparency and explainability in AI systems, ensuring users are aware of when they are interacting with AI and how it functions. The goal of preventing deception and fostering informed user consent is a universal concern in responsible AI development, making Washington's approach consistent with international best practices for trustworthy AI.

Furthermore, SB 5984's focus on protecting vulnerable populations, especially minors, and addressing mental health risks, such as suicidal ideation, reflects a growing global consensus on the need for human-centric AI regulation. International bodies and national governments are increasingly recognizing the specific harms that AI, particularly emotionally engaging chatbots, can inflict on children and adolescents. The prohibitions against sexually explicit content and emotionally manipulative techniques for minors, as well as the mandate for suicide prevention protocols, mirror similar safeguards being explored or implemented in other jurisdictions. While the specific legal mechanisms may differ, the underlying ethical imperatives—to prioritize human well-being, prevent exploitation, and mitigate psychological harm—are shared across diverse regulatory landscapes. This alignment suggests that despite being a sub-national regulation, Washington SB 5984 contributes to the global discourse on responsible AI and demonstrates a localized application of universally recognized ethical principles in AI governance.

Implementation Timeline

Compliance Checklist

Sources and References