Washington AI Law

This page summarizes selected Washington statutes concerning AI and synthetic media, distinguishes enacted laws by their effective dates, and preserves a historical legislative tracker. The summaries are not a complete inventory of Washington law applicable to AI systems.

Enacted laws

Effective in 2027

AI content provenance and government chatbot disclosure — E2SHB 1170

E2SHB 1170 was signed on March 24, 2026, as Chapter 167, Laws of 2026. It takes effect February 1, 2027.

Covered providers produce generative AI systems with more than one million monthly users that are publicly available in Washington for personal use; governments are excluded. Subject to commercial and technical reasonableness, providers must embed provenance information in generated or materially altered images, audio and video and protect it against removal or tampering. Minor edits are excluded from material alteration. Exceptions cover business-to-business uses, exclusively video-game or interactive experiences, and systems used solely for upscaling, noise reduction or compression. Enforcement under this chapter's Consumer Protection Act provision is reserved to the Attorney General.

Separately, government agencies must disclose consumer interactions with AI before or when interaction begins. The notice must be conspicuous, use plain language and avoid dark patterns; a link to another page may provide it. This obligation applies even when the system's artificial nature would be obvious.

AI companion chatbots — ESHB 2225

ESHB 2225 was signed on March 24, 2026, as Chapter 168, Laws of 2026. It takes effect January 1, 2027.

Operators must disclose that a companion chatbot is artificial at the outset and at least every three hours of continued interaction, and reasonably prevent claims of being human. For known minors or systems directed to minors, reminders are required at least hourly, alongside reasonable measures against sexual content, suggestive dialogue and specified emotional-manipulation techniques. Operators must implement self-harm detection and response protocols, offer crisis referrals, take reasonable measures against encouragement or instructions for self-harm, and publish protocol details and annual referral counts. Violations fall under the Consumer Protection Act.

Exclusions have distinct conditions: business tools and voice-assistant devices have relationship and emotional-response limits; game bots must remain within game-related topics and exclude mental-health, self-harm and sexual discussions; narrowly tailored educational tools must serve curriculum objectives without open-ended companionship. Section 2 supplies the full definitions.

In effect

Fabricated sexually explicit depictions of minors — 2ESSB 5105

2ESSB 5105 was signed on March 25, 2026, as Chapter 223, Laws of 2026, and took effect June 11, 2026.

The law expands Washington criminal provisions concerning fabricated depictions of minors engaged in sexually explicit conduct. Among other changes, it extends the definition of a fabricated depiction to specified obscene material, changes defenses involving whether a depicted minor exists or is identifiable, increases the limitations period for specified depiction offenses from three years to ten years, and expands the offense of sexual exploitation of a minor to specified conduct involving an unconscious or unaware minor.

Forged digital likenesses and personality rights — SSB 5886

SSB 5886 was signed on March 16, 2026, as Chapter 69, Laws of 2026, and took effect June 11, 2026.

The law extends Washington's personality-rights statutes to a person's forged digital likeness. It defines that term to cover specified persistent or real-time visual representations and audio recordings of an identifiable person that may be indistinguishable from genuine material, misrepresent the person's appearance, speech, or conduct, and are likely to deceive a reasonable person. Unauthorized covered uses may support injunctive relief, a $3,000 civil penalty, actual damages, attributable profits not already counted in actual damages, and—in forged-digital-likeness cases—noneconomic damages whether or not the infringement produced a profit.

Disclosing fabricated intimate images

RCW 9A.86.030, enacted through Chapter 88, Laws of 2024, addresses knowingly disclosing fabricated intimate images of an identifiable person when the discloser knows or should know of nonconsent and knows or reasonably should know that disclosure would cause harm. Its definition includes realistic computer-generated or altered images, whether produced manually or automatically, and expressly includes AI. A different standard applies to defendants under eighteen: intentional and malicious disclosure and knowledge or constructive knowledge of nonconsent.

The statute excludes specified public-interest disclosures and constitutionally protected commentary, criticism or disclosure, and limits liability for service providers based solely on another person’s content. Consent to creating or altering an image is not a defense to the disclosure offense. The first offense is a gross misdemeanor; a prior conviction under this section or the related intimate-image disclosure provision makes it a class C felony. Section .020 provides for forfeiture of covered images possessed by minors. These criminal provisions are distinct from available civil remedies. See deepfake topics.

Civil remedies for fabricated intimate images

RCW 7.110.025 gives an identifiable depicted individual who suffers harm a civil claim for intentional disclosure or threatened disclosure of a fabricated intimate image without consent, where the defendant knew or recklessly disregarded nonconsent and identifiability. Creation consent alone does not establish disclosure consent; valid consent requires a knowingly and voluntarily signed, plain-language agreement describing the image. An unauthorized-image disclaimer is not a defense.

Section .030 supplies exceptions, including specified good-faith disclosures, public-interest material and constitutionally protected expression. Public-figure status alone does not establish public interest; the parental exception applies to the separate .020 intimate-image claim.

Remedies include the greater of compensatory damages or statutory damages capped at $10,000 per defendant across covered disclosures, attributable gain, and punitive damages up to three times the damages under .050(1)(a). Fees and injunctions may also be awarded. Disclosure claims generally have a four-year discovery period; threats have four years from the threat, with tolling provisions and delayed commencement for minors. Section .901 preserves consistency with 47 U.S.C. §230. See deepfake topics.

Government facial-recognition systems

Chapter 43.386 RCW regulates state and local government use of facial-recognition services. It addresses public notice and accountability reports, meaningful human review, testing and mitigation of performance differences, training, disclosure to criminal defendants, record retention, surveillance uses, and statutory exemptions. Most provisions took effect July 1, 2021.

Synthetic media in electioneering communications

Chapter 29B.35 RCW permits a candidate whose appearance, action, or speech is altered through synthetic media in an electioneering communication to seek injunctive or other equitable relief and, in specified circumstances, damages against the sponsor. The chapter provides a disclosure-based affirmative defense and rules concerning the liability of media that disseminate such communications.

RCW 29B.35.010 places the burden of proving synthetic-media use on the plaintiff by clear and convincing evidence. The disclosure defense specifies prominent visual text or clearly audible notices at the beginning and end and, for longer audio, at intervals of no more than two minutes. Section .020 generally places liability on the sponsor, with specified exceptions for a medium that removes disclosures or changes content into synthetic media. Section .030 requires Public Disclosure Commission rulemaking but does not make the chapter a violation of other campaign-finance provisions or independently authorize Commission action under RCW 29B.60.020.

AI task-force framework — 2024 Chapter 163

Engrossed Second Substitute Senate Bill 5838 became Chapter 163, Laws of 2024, on March 18, 2024, with immediate effect. Subject to appropriations, section 2 establishes an Attorney General-administered task force to assess AI uses and recommend guidelines and possible legislation. It provides for legislative, government, industry, labor and community representation and advisory subcommittees.

Its statutory work covers existing legal protections, discrimination, high-risk uses, training data, transparency, personhood and likeness protections, employment tools, human oversight and potential remedies. Section 2(5) sets preliminary, interim and final report deadlines of December 31, 2024, December 1, 2025 and July 1, 2026; section 2(9) provides for that section’s expiration on June 30, 2027. Recommendations do not themselves enact the proposed policies.

The Attorney General’s current index lists the final report as released July 1, 2026, alongside two earlier reports and three commissioned studies. The index confirms publication, rather than legislative adoption of the recommendations. See governance topics.

Historical tracker

May 1, 2026 tracker snapshot

The following entries preserve the historical NCSL snapshot used by the former state directory. Status labels such as “enacted” and “failed” describe that snapshot rather than a current-status assessment. This tracker is not a complete inventory of enacted Washington law.

NCSL's May 1, 2026 snapshot listed 19 AI-related measures for Washington.

Enacted measures recorded in the snapshot

• WA H 1170 — AI content provenance and disclosures; enacted. Topics: Notification, Private Sector Use, Responsible Use.
• WA H 2225 — AI companion chatbot regulation; enacted. Topics: Health Use, Notification, Private Sector Use.
• WA S 5105 — Fabricated sexually explicit depictions of minors; enacted. Topics: Child Pornography, Criminal Use, Deepfake.
• WA S 5886 — Forged digital likenesses and personality rights; enacted. Topics: Deepfake.

Other measures recorded in the snapshot

• WA H 1168 — Transparency in artificial intelligence; failed. Topics: Government Use, Notification, Oversight/Governance, Private Sector Use, Responsible Use.
• WA H 1622 — Bargaining over matters related to AI use; failed. Topics: Government Use, Effect on Labor/Employment.
• WA H 1833 — Artificial intelligence grant program; failed. Topics: Appropriations, Health Use, Effect on Labor/Employment, Oversight/Governance.
• WA H 2157 — High-risk AI systems; failed. Topics: Health Use, Impact Assessment, Private Sector Use, Private Right of Action.
• WA H 2503 — AI training data; failed. Topics: Deepfake, Provenance.
• WA H 2667 — Consumer protections for AI systems; failed. Topics: Government Use, Health Use, Impact Assessment, Private Sector Use, Responsible Use, Studies.
• WA S 5469 — Algorithmic rent fixing and rental-housing noncompete agreements; failed. Topics: Housing, Private Sector Use.
• WA S 5637 — Media literacy and civic education; failed. Topics: Deepfake.
• WA S 5870 — Civil liability for suicide linked to AI systems; failed. Topics: Health Use, Notification, Oversight/Governance, Private Sector Use, Private Right of Action.
• WA S 5956 — AI, student discipline, and surveillance in public schools; failed. Topics: Education Use.
• WA S 5984 — AI companion chatbot regulation; failed. Topics: Health Use, Notification, Private Sector Use.
• WA S 6120 — High-risk AI systems; failed. Topics: Health Use, Impact Assessment, Private Sector Use, Private Right of Action, Responsible Use.
• WA S 6254 — AI-assisted streamlining of rules and regulatory guidance; failed. Topics: Government Use.
• WA S 6284 — Consumer protections for AI systems; failed. Topics: Government Use, Health Use, Impact Assessment, Notification, Private Sector Use, Responsible Use, Studies.
• WA S 6299 — AI and instructional staff; failed. Topics: Education Use.

Seattle municipal ordinance (2026)

Seattle council bill on algorithmic grocery pricing (2026)

The Seattle City Council passed CB 121267 on September 22, 2026. Mayor Katie Wilson signed it on October 5, and the City Clerk attested it as Ordinance 127515. Section 1, which adds Seattle Municipal Code Chapter 7.35, takes effect September 1, 2027.

The ordinance prohibits covered grocery businesses, mixed-use grocery businesses and delivery providers from using automated analysis of consumers' personal information to vary specified prices. Coverage varies by retailer type; the text provides exceptions for specified delivery-cost differences, generally available discounts and defined loyalty-program tiers. It also requires price and discount disclosures and compliance records, with enforcement by the City Attorney and a limited private action.

See September 24 council-passage coverage and October 6 coverage of the mayoral signature.

2026 coverage

• March 13, 2026 digest — Washington coverage, with subsequent enactment details