08/04/2026
Every state legislature is deciding AI policy this year. Most are doing it the way legislation normally gets made: a bill drafted by a small number of people, introduced, amended under public pressure, voted on within weeks.
Colorado tried that first. The state's 2024 AI law drew immediate pushback, and two legislative sessions, including a special session, failed to fix it.
In 2025, Colorado tried something different. A fourteen-member working group spent nine months getting the definitions right before a bill was written. The result passed the Senate 34–1 and the House 57–6, signed thirteen days after introduction.
We designed and facilitated that process. Our new white paper, Getting AI Regulation Right, sets out the mechanism in enough detail to be usable by a legislator, a governor's office, or an attorney general's office considering the same approach, and walks through eight domains where states are already deciding AI policy: consequential decisions, mental health and minors, frontier AI safety, public safety, data centers, workforce impact, government's own AI use, and synthetic media.
Read it here:
Most AI legislation gets written the way legislation normally gets made: a bill arrives already drafted, and every party’s first move is to defend or attack a position instead of understanding one. This paper makes the case for a different approach: get the people with real expertise in the room, ...