HB1170 creates a new chapter in Washington law requiring certain publicly accessible providers of generative artificial intelligence systems to disclose when image, video, or audio content has been created or altered by AI. The bill defines key terms such as artificial intelligence, generative AI system, provenance data, system provenance data, personal provenance data, and “covered provider,” which is generally a provider of a publicly accessible generative AI system with more than 1 million monthly users in the state. It also excludes products or services used exclusively for video game, television, streaming movie, or interactive experiences.
The bill requires covered providers to offer a free AI detection tool that can assess whether content was created or altered by the provider’s generative AI system, detect and output system provenance data, avoid outputting personal provenance data, be publicly accessible, and allow uploads or URL-based checks, including API access. Covered providers must also collect user feedback on the tool and use it to improve efficacy, while limiting collection and retention of user personal information and submitted content.
In addition, the bill requires covered providers to offer users the option to include a visible “manifest” disclosure in AI-generated or AI-altered content and to include a latent disclosure that is technically feasible, permanent or difficult to remove, and detectable by the provider’s detection tool. For licensed systems, providers must contractually require third-party licensees to preserve disclosure capability, and if a licensee modifies the system so disclosures can no longer be included, the provider must revoke the license within 24 hours of discovering the issue. The bill also states that violations are unfair or deceptive acts and unfair methods of competition under the Consumer Protection Act, and it creates a new chapter in Title 19 RCW.
The bill’s impact is to impose new disclosure, detection, and licensing obligations on large generative AI providers operating in Washington, while also creating a consumer-protection enforcement hook through the state’s Consumer Protection Act. It would likely affect AI platform operators, developers, licensors, and third-party licensees, especially those distributing image, video, or audio generation tools to Washington users. The bill does not apply to certain entertainment-focused services, narrowing its reach.
Overall, the sentiment appears generally favorable but not unanimous. The bill advanced through both chambers with multiple committee approvals and ultimately passed the House and Senate, though several House votes were relatively close, suggesting meaningful concern or disagreement. Senate votes were much stronger, including unanimous committee approval and a 46-3 floor vote, indicating broader support after amendment.
The main points of contention likely center on the burden placed on AI companies, the technical feasibility of permanent or detectable disclosures, privacy implications of detection tools, and the scope of the covered-provider definition. Supporters appear to favor transparency, consumer protection, and provenance tracking for AI-generated media, while opponents or skeptics likely worry about compliance costs, implementation complexity, and whether the requirements are workable for rapidly evolving AI systems.
HB1170 would add a new chapter to Title 19 RCW and make AI-content disclosure and detection requirements part of Washington consumer protection law. It would regulate covered providers of generative AI systems by requiring free detection tools, disclosure options, provenance-preserving features, and contractual controls over third-party licensees. Violations would be treated as unfair or deceptive acts and unfair methods of competition under the Consumer Protection Act, giving the state enforcement leverage and potentially exposing covered providers to consumer-protection remedies.
The bill appears to have broad but not unanimous support. It moved through House and Senate committees and passed both chambers, with especially strong support in the Senate after amendments. However, the House floor votes were more divided, indicating that while many lawmakers supported AI transparency measures, a substantial minority had reservations about the bill’s approach or scope.
The likely areas of contention are the compliance burden on large AI providers, the technical feasibility of creating permanent or difficult-to-remove disclosures, and the privacy implications of requiring detection tools that process user-submitted content. Another point of debate is the bill’s scope: it targets large publicly accessible generative AI systems but excludes certain entertainment products and services, which may have raised questions about fairness, coverage, and whether the definition of covered provider is too broad or too narrow. The licensing and revocation provisions for third-party users of AI systems may also have been controversial because they impose downstream contractual and operational obligations.