WA HB2667

Requires deployers of high-risk AI systems (those making consequential decisions about parole, education, employment, lending, housing, healthcare, insurance, or legal services) to implement risk management programs and complete impact assessments beginning July 1, 2027

Defines "algorithmic discrimination" as unlawful differential treatment based on protected classes under Washington's anti-discrimination law (RCW 49.60) or federal law, and requires deployers to conduct annual reviews to ensure AI systems are not causing such discrimination

Mandates consumer notification before AI systems make or substantially factor into consequential decisions, including disclosure of the system's purpose and plain-language description, effective July 1, 2026

Requires government agencies using AI systems to interact with consumers to clearly disclose AI involvement before or at the time of interaction, without using dark patterns

Extends the existing AI task force through June 30, 2028 and creates a new AI workplace advisory group to develop guiding principles for ethical AI use in employment settings, with reports due December 2026 and March 2027