Rep. Stephanie Hu (D-50) introduced HB 2843 in the Washington House today, a companion to last session's HB 1170 that addresses what advocates have called the "decision laundering" loophole. Where HB 1170 required employers to disclose the use of AI in employment decisions, HB 2843 requires that a named human being review and personally take responsibility for any adverse outcome the AI recommends.
The bill defines "adverse outcome" as a denied hire, a denied promotion, a denied raise, a denied accommodation, a negative performance review, a denied leave request, or a termination. The named reviewer must read the AI's recommendation and the underlying inputs, document that they did so, and sign their name. Civil penalties begin at $5,000 per violation. Workers gain the right to request the reviewer's name and the AI's stated reasoning.
"The phrase 'AI made the decision' has become a way to make sure nobody made the decision," Hu said. "We're not banning AI in HR. We're saying: if it tells someone they don't get the job, a person has to sign their name to that. Not a department. A person."
The bill is co-sponsored by Reps. Reeves (D-43), Okafor (D-37), Murakami (R-12), Lopez (D-11), and Stein (D-46). It has been referred to the House Innovation & Technology Committee, where Hu sits.
Care-worker and labor groups in the state have signaled support; technology-industry groups have not yet taken a position. A Senate companion is expected later this session.
Press inquiries: press@repstephaniehu.com