Washington State AI Task Force Final Report: Status of 11 Policy Recommendations

Washington — From Christine Brewer, PIA Washington Lobbyist: Brewer Public Affairs Interim Update: WA State AGO's AI Task Force Final Report

This interim update is to flag for your interest the release of the Attorney General’s Office Artificial Intelligence Task Force final report last month.

Washington's AI Task Force (created by SB 5838 (2024)) released its final report on July 1, wrapping up two years of work. Over that span it advanced 11 policy recommendations to the Legislature: 4 contained either directly or aligned with bills already enacted this session, 4 concepts that had bills introduced but stalled, and 3 still without a legislative vehicle. Two new recommendations (a permanent AI advisory body, and companion chatbot regulation) were adopted at the Task Force's final vote in April.

Big Picture

Attorney General Nick Brown's framing is "innovation vs. protection isn't a binary." The report uses the NIST AI Risk Management Framework (and ISO/IEC 42001) as the baseline governance model, pushes a risk-based approach that scales oversight to high-risk uses (health care, employment, housing, lending, criminal justice), and focuses on algorithmic discrimination, transparency, and training-data disclosure as the core themes.

Task Force recommendations already enacted

Companion AI chatbot rules, disclosure + minor safeguards (ESHB 2225)

Healthcare prior-auth transparency, AI can't be sole basis for care denial (E2SSB 5395)

Law enforcement AI disclosure, enacted in part, limited to license plate readers (ESSB 6002)

AI-generated CSAM liability fix (2ESSB 5105)

Recommendations that had bills, but stalled/died

High-risk AI governance mandate (NIST/ISO framework + public disclosure) (HB 2157 and SB 6284)

AI development/training-data transparency (HB 1168)

AI innovation grant program, HB 1833 (passed House, died in Senate Ways & Means)

Worker-centered AI workplace guidelines, committee amendment to HB 1833, rejected

No bill yet

Permanent AI/emerging-tech advisory body (new)

K-12 STEM/AI literacy investment (budget issue)

Formal adoption of NIST ethical AI principles

Going Forward

Rep. Cindy Ryu's HB 2157 is important to watch for industry as it would require both developers and deployers of "high-risk" AI (hiring, insurance, lending, housing, school admissions, etc.) to use "reasonable care" against algorithmic discrimination, with enforcement via court-ordered relief (no damages) rather than a full private right of action. Sen. Marko Liias introduced a Senate vehicle on the same theme (SB 6284). Neither passed this year, but both align with Task Force recommendations, and will likely return in 2027. Rep. Ryu has left the House to run for the Senate, so leadership on the issue in both chambers may change next session.

Rep. Clyde Shavers remains one of the most active legislators in this space. Beyond HB 1168 above, he also sponsored HB 1170, a separate AI content-detection/deepfake-disclosure bill that was signed into law this session, although it wasn't itself a Task Force recommendation. This summer he is stakeholdering a draft AI liability/sandbox bill for likely introduction in '27. The proposal in its current form would amend Washington's product liability statute (RCW 7.72) to give AI developers and deployers a rebuttable liability shield if they document risk testing, disclose known risks, and, for products likely to reach minors, build in age/content safeguards. The draft would pair that with a new Commerce Department "AI regulatory sandbox" letting companies test systems for up to 18 months with relief from certain regulatory requirements, subject to quarterly reporting. Expect to see some form of this proposal in the pre-filed bills come December.