Starting February 1, 2027, large AI providers must attach origin information to covered media when reasonable, and agencies must tell consumers when they interact with AI.
Large providers must attach information showing when their AI creates or significantly changes audio, images, or video, when doing so is reasonable. Minor edits and some products and uses are excluded.
Agencies must tell consumers before or when they interact with an AI system, even if the interaction seems obvious. The clear, plain-language notice may link to another webpage.
Covered providers must include information showing whether their AI created or significantly changed covered media. This applies to providers of publicly available generative AI systems with more than 1 million monthly users in Washington. State, local, and tribal governments are excluded from this provider definition. Providers must include the information only when doing so is commercially and technically reasonable. They must use commercially and technically reasonable methods to make the information hard to remove or tamper with. Using a commonly supported watermarking or metadata standard counts as compliant. The information need not identify a person. Minor edits that do not significantly change how the content looks or sounds, such as cropping or removing background noise, do not count as significant changes. The chapter does not cover business-to-business uses or sales, video games or interactive shopping experiences, or systems used only for upscaling, noise reduction, or compression. The law takes effect February 1, 2027.
Government agencies must tell consumers when they are interacting with an AI system. This applies when a government agency makes available an AI system meant to interact with consumers. The notice must be clear, easy to see, and written in plain language. It cannot use a dark pattern, or design that steers people without making that choice clear. An agency may use a link to a separate webpage for the notice. It must give notice even if a reasonable consumer would know they are interacting with AI. The law takes effect February 1, 2027.
No hearings or votes scheduled.