AN ACT Relating to regulating high-risk artificial intelligence system development, deployment, and use;
HB 2157 creates a new chapter in Title 19 RCW regulating the development, deployment, and use of “high-risk artificial intelligence systems” in Washington. The bill defines a wide range of AI-related terms, including algorithmic discrimination, consequential decisions, deployers, developers, facial recognition, generative AI, synthetic content, and impact assessments. It focuses on AI systems used in high-stakes contexts such as parole, education, employment, lending, healthcare, housing, insurance, marital status, and legal services.
The bill requires developers of high-risk AI systems to use reasonable care to protect consumers from known or reasonably foreseeable risks of algorithmic discrimination, and it requires developers to provide documentation and disclosures about intended uses, limitations, performance, and mitigation measures. Deployers must adopt risk management policies, complete pre-deployment and post-deployment impact assessments, disclose to consumers when they are interacting with AI, provide notices and explanations for adverse consequential decisions, and maintain records for several years. The bill also imposes special disclosure and labeling requirements for certain generative AI outputs, including synthetic content identification, subject to exceptions for artistic, satirical, public-interest, and assistive uses.
The bill would add a new consumer-protection framework to Washington law governing AI systems, creating affirmative duties for both AI developers and deployers and authorizing civil actions for violations. It would require documentation, disclosures, risk management programs, impact assessments, consumer notices, and record retention, while also creating exemptions for certain regulated sectors and activities, including some financial institutions, insurers, federal systems, research, security, and public-interest uses. The measure would likely affect technology companies, employers, lenders, housing providers, insurers, and other entities using AI in consequential decision-making, and it would interact with existing state and federal anti-discrimination, privacy, and consumer protection laws.
The available vote history suggests the bill had meaningful but not unanimous support in committee, passing the House Technology, Economic Development, & Veterans Committee 8-5 on a substitute bill. That vote pattern indicates general interest in regulating high-risk AI, but also substantial concern about the scope and burden of the proposal. No committee transcript was provided, so the record here reflects the committee vote rather than detailed debate. Overall, the bill appears to have been viewed as a consumer-protection and civil-rights measure, with support from members favoring oversight and skepticism from members concerned about regulatory reach.
The main points of contention are likely the breadth of the bill’s definitions and compliance obligations, especially for businesses deploying AI in consequential decisions. Potential critics may object to the costs and operational complexity of risk management policies, impact assessments, disclosures, and recordkeeping, as well as the possibility of litigation and uncertainty around what counts as algorithmic discrimination or a high-risk system. Supporters are likely to emphasize consumer transparency, anti-discrimination protections, and accountability for AI-driven decisions. The bill also contains notable carve-outs and exemptions for federal systems, certain financial and insurance activities, research, security, and protected speech, which suggests lawmakers were balancing consumer protections against concerns about overregulation and interference with legitimate uses of AI.