Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai Policy Lab topic
No spam. Unsubscribe anytime.
Utah’s AI policy lab tests targeted 'mitigation agreements' instead of model-level bans
Summary
Utah lawmakers and industry panelists described the state’s AI policy lab, which signs use-case-specific mitigation agreements with companies to pilot AI deployments, learn from them and inform narrowly tailored regulation rather than adopting broad model-level rules.
Utah has favored a laboratory approach to regulating artificial intelligence, creating an AI policy lab that negotiates use-case-specific mitigation agreements with companies instead of imposing prescriptive, model-level rules.
Sen. Kurt Colmore, who described the lab’s origins, said the state convened an informal working group in 2023 with legislators, the Department of Commerce, technology services, academia and practitioners to consider how to address consumer-facing AI. "We kind of flipped it on its head a little bit and said, let's do an AI policy lab," Colmore said, explaining the lab can "enter into what we call mitigation agreements" that allow limited pilots, liability adjustments or exemptions tied to specific uses.
The lab is intended to let Utah test deployments that might otherwise operate in regulatory "gray areas," learn from those pilots and then recommend narrower, targeted rules to regulators and the legislature. Colmore said the approach has been in place "about a year and a half." He framed the lab as a way to protect consumers without "stifling innovation."
Industry supporters on the panel said the lab approach helped craft narrowly tailored safeguards. Chris Koopman, CEO of the Abundance Institute, called AI a "general purpose technology" and said Utah’s focus on use-case regulation is preferable to omnibus model-level rules that risk inhibiting unexpected beneficial uses. Koopman cited a mental-health chatbot pilot that emerged from the lab as an example of a narrowly scoped effort to understand risks and appropriate guardrails.
Panelists contrasted Utah’s consumer-facing, use-case regulatory focus with other states that have considered or adopted model-level regulation. They urged continued iterative learning: use the lab to surface real-world harms, adjust mitigation agreements and then codify duties or limits where necessary.
The panel concluded that state labs can inform future regulation while keeping development and entrepreneurship active; speakers urged ongoing stakeholder engagement so pilots inform lawmaking rather than create surprise disruptions for businesses.
Next steps: panelists said they expect the lab to continue negotiating pilots and to provide legislators with empirical findings that could shape future, narrowly tailored rules.
