Protect

Protect Coloradans and the Environment

The opportunity

77%

of voters, Republican and Democrat alike, prefer a candidate who protects consumers and kids from AI harms (Transparency Coalition, 2026).

Coloradans want the confidence that AI is on their side, and protection is how they get it. A leader who holds companies accountable when their AI harms people earns Coloradans' trust, and their vote.

Colorado already holds people accountable for these harms, and can hold AI to the same standard

Each harm below is already a wrong the law addresses when a person causes it. When AI causes one, both the person who misused it and the company that built it answer.

AI-generated child sexual abuse material
Non-consensual sexual deepfakes
Chatbot manipulation preceding self-harm
AI companions exploiting emotional dependency for commercial gain
Automated denial of essential care or benefits
Algorithmic price-fixing that raises what people pay

A new private right of action adds the missing claim against the company that built the system.
The attorney general can enforce too, both powers limited to these named harms and no further.
Naming the harms keeps the law specific and prosecutable.

90%

of Americans say families should be able to sue an AI company that harms their child (YouGov, 2025)

Colorado already gives good businesses clear rules, and can give AI builders the same

Good actors get clear protection. A builder who adopts the NIST AI Risk Management Framework, publishes its safety practices, and reports harms earns an affirmative defense against the named-harm lawsuits.

Built on the NIST AI Risk Management Framework, the closest thing the country has to a recognized AI safety standard, with no licensing, pre-approval, or upfront paperwork.
The affirmative defense covers every named harm, CSAM and non-consensual deepfakes included, and is earned by proving good-faith adherence.
A first-time slip can be cured before any penalty, but a builder that ignores the framework gets no shield and can be sued, with the court deciding liability.

80%

of voters, across both parties, want enforceable AI safety standards over letting companies regulate themselves (AI Policy Institute, 2024)

Colorado already protects its water and renewable energy, and can hold AI data centers to the same standard

There is no statewide rule today, so cities are banning data centers one at a time. One clear standard gives builders certainty: where they can build and what they must meet.

Bring your own new renewable power
No potable water for cooling
Public reporting of energy and water use

As electricity demand climbs more than 40% by 2035 (Xcel), builders that bring their own new power add supply to the grid instead of straining it.

78%

of Americans worry new data centers will raise their energy bills (Consumer Reports, 2025)

What this creates

87%

of Americans say AI companies should answer for the harms they cause (Fathom, 2026).

Be the first state to make AI answer to the people it affects

Built to fit the child-safety, fraud, and consumer-protection powers the March 20, 2026 federal framework leaves to the states, so the protections hold when a Coloradan is hurt.

The ask

We ask Colorado's next governor, attorney general, and senator to endorse the movement and champion these protections.

Check the ones you want your next leaders to deliver. Your picks come with you when you join.

For named and known harms only caused by AI companies, establish a private right of action against the AI companies as well as allow attorney general enforcement, specifically limited to the named harms, with penalties matched to the harmThis simply extends what is already illegal to AI companies that have caused harm.Create a safe harbor for good actors who adopt the NIST AI Risk Management Framework and publish their safety practices, with a right to cure and protection from the named-harm lawsuits when they followed best practicesNo such safe harbor exists today, so create and sign it into law by the end of 2027.Set renewable-energy and water-recycling standards for AI data centers, so builders know the rules and can buildNo such standards exist today, so create them by the end of 2027.

0 of 3 on this pillar,0 of 10 pledges across the framework.

When Colorado protects Coloradans from AI's harms, it earns the trust to govern and lead with it.

See the full ten-pledge scorecard

Sources

Voter preference for protecting people from AI (77%, Republicans and Democrats alike): Transparency Coalition, 2026. Republican crossover (71% prefer a candidate who supports regulating AI): AI Policy Institute, 2024. Corroborated by Fox News (nearly 80% call government action on AI urgent, 2026) and Gallup (97% say AI should be subject to rules, 2025).
Support for a family's right to sue over AI harm to a child (90%): YouGov for the Institute for Family Studies, 2025.
Support for holding AI companies accountable for harms (87%): Fathom national survey, fielded by Forbes Tate Partners, 2026.
Voter preference for enforceable AI safety standards over industry self-regulation (80%): AI Policy Institute national poll, June 2024 (76% of Republicans, 84% of Democrats; only 24% prefer industry self-regulation).
Projected peak electricity demand, 2021 to 2035: Xcel Energy. Concern that new data centers will raise energy bills (78%): Consumer Reports, 2025.
Legal basis for the named harms: federal PROTECT Act (CSAM), TAKE IT DOWN Act (deepfakes), tort and consumer-protection law (chatbot and companion harms), insurance bad-faith law (claim denial), and the Sherman Act (price-fixing). The named harms are Colorado AI framework proposals. Colorado's SB 25-288 (2025) added AI-generated CSAM to state law for an identifiable child, and the framework recommends aligning Colorado's definition to the federal PROTECT Act standard so it also reaches fully synthetic CSAM, a clean child-safety fix.
Federal framework: EO 14365 (Dec 11, 2025) and the March 20, 2026 National AI Policy Framework, which preserves state police powers over child safety, fraud, and consumer protection.