Washington AI Task Force Ends With Narrow Laws, Broader Debates

Washington state's two-year AI task force has concluded its work, but the fundamental tension between innovation and regulation remains unresolved. While the group successfully shepherded four AI laws through the legislature this spring, the measures that passed were narrowly focused—and the broader, more consequential proposals died without reaching a vote.

The contrast reveals the political challenges facing AI regulation at the state level, particularly in a tech-heavy state caught between industry concerns about compliance costs and demands from consumer advocates for stronger safeguards.

What passed and what didn't

The four laws that became reality address specific, bounded use cases: companion chatbots must disclose they're not human, and medical insurers cannot deny coverage based solely on AI assessments. Two other sector-specific measures also cleared the legislature.

What failed tells a different story. A bill to regulate AI in high-risk decision-making—hiring, pricing, criminal justice, healthcare—never reached a floor vote in either chamber, despite similar laws passing in New York, Connecticut, Illinois, California, and Colorado. Requirements for AI developers to disclose training datasets, as California mandates, also stalled. So did workplace AI guidelines.

According to GeekWire, which first reported these details, Yuki Ishizuka of the Washington Attorney General's Office attributed the pattern to political dynamics: "What passed in the legislature was more sector specific things where the consumer harm was more clear. It's harder to connect broader governance or transparency bills to harm to people."

State Rep. Mia Gregerson, who sponsored comprehensive bills, told GeekWire that Washington has "an even bigger responsibility to do good work to catch up to what other states are doing. We are so behind."

The startup compliance question

At a panel discussion Wednesday at Seattle's AI House incubator, startup founders raised concerns about regulatory burden. The event highlighted a recurring theme in AI policy debates: whether small companies should face the same compliance requirements as Meta or Google.

University of Washington law professor Ryan Calo, speaking at the panel, offered a pragmatic view of enforcement priorities. He noted that even the narrow laws passed carry significant teeth—violations can establish negligence per se and trigger Washington's Consumer Protection Act. But he suggested enforcement would likely focus on major players: "If you're a little startup, probably not, but if you're Meta, probably yes."

Jai Jaisimha of the Transparency Coalition, himself a former startup founder, pushed back against carve-outs based on company size, arguing that "disclosure and consumer protection, these are standard practices in other industries."

Federal pressure and what's next

The Trump administration's December executive order threatens to cut broadband funding to states with AI regulations that don't align with federal priorities, specifically naming Colorado. The order established an AI Litigation Task Force to challenge state laws and called for Congress to preempt state regulations.

The pressure appears effective: Colorado repealed and replaced its comprehensive 2024 AI law this spring. Utah withdrew a frontier model bill after federal criticism, according to Politico.

Washington's Attorney General's Office signaled it will continue AI policy work through a new Tech Policy Team led by Ishizuka, despite the task force's dissolution. Rep. Gregerson told GeekWire she plans to pursue funding for retraining workers displaced by autonomous vehicles. Rep. Clyde Shavers, a task force member, has indicated interest in establishing liability frameworks for AI-related harm.

Why it matters

Washington's experience demonstrates that even states with strong tech sectors and political will struggle to enact comprehensive AI regulation. The gap between what passed and what failed reveals how difficult it is to regulate general-purpose technology when concrete harms are hard to illustrate and industry opposition is well-resourced. With federal preemption threats escalating and no congressional AI framework in sight, states face growing uncertainty about whether their regulatory efforts will survive legal challenge—or whether they should attempt them at all.

These details were first reported by Grace Kaste for GeekWire.