A new, powerful Citizen Portal experience is ready. Switch now

Colorado committee considers bill to require explainable AI for consequential state decisions

February 26, 2026 | 2026 Legislature CO, Colorado


This article was created by AI summarizing key points discussed. AI makes mistakes, so for full details and context, please refer to the video of the full meeting. Please report any errors so we can fix them. Report an error »

Colorado committee considers bill to require explainable AI for consequential state decisions
Representative Paschal proposed a draft bill that would require any Colorado state department that selects software using artificial intelligence to make a "consequential decision" to demonstrate two features before deployment: that the system can explain how a decision was made, including the data used, and that it "consistently comes to the same decision every time when using the same set of data." Paschal said the proposal builds on SB 24-205 and is targeted at state agency usage.

"Before any department of the Colorado government can select software containing I.E.I. to make a consequential decision it must first demonstrate that the software, can explain how the decision was made including the data that went into making the decision and can ... demonstrate that it consistently comes to the same decision every time when using the same set of data," Representative Paschal said.

Legislative Council staff member Samantha Falco explained the drafting process: the committee would vote to have a bill drafted and then vote to introduce the draft, and staff from the Office of Legislative Legal Services were available to assist. Falco also said fiscal staff would flag AI-related bills going forward to help the committee track other sponsors and proposals.

Several committee members urged stakeholder outreach before finalizing text. Representative Weinberg, who said he has worked on related issues professionally, recommended an overview meeting with industry and raised two verification concerns: "who's gonna verify the data, and then who's gonna verify the person who's verifying the data?" Senator Rodriguez suggested inviting the Attorney General's office because the AG is likely to be the enforcer for any verification regime.

The chair read a set of standard questions the committee plans to send to state agencies to understand current AI use, including whether agencies use systems meeting the "high risk AI" definition, whether AI has replaced or is slated to replace staff positions, the vendors contracted, risk-management frameworks, where Colorado residents' biometric and personal data are stored, retention policies, and whether AI tools are used to determine distribution of services or funds. Rep. Paschal stressed many of those questions must prompt substantive answers beyond yes/no responses.

The committee agreed to revise the questions for clarity, allow members about a week to propose edits, and then send them to agencies with an expectation of responses within two weeks after distribution. The chair proposed, and members concurred, that the Attorney General's office and the Office of Information Technology (OIT) should appear next week to discuss enforcement, verification, and how oversight might work in practice. Representative Paschal and others were asked to begin stakeholder outreach in the interim.

Next steps: the committee will refine the agency questionnaire, request written responses from state agencies, and host the AG and OIT to discuss enforcement and technical implementation before deciding whether to put a bill draft on paper.

View the Full Meeting & All Its Details

This article offers just a summary. Unlock complete video, transcripts, and insights as a Founder Member.

Watch full, unedited meeting videos
Search every word spoken in unlimited transcripts
AI summaries & real-time alerts (all government levels)
Permanent access to expanding government content
Access Full Meeting

30-day money-back guarantee