Colorado’s AI Act Rewrite Moves Forward: What Employers Need to Know
Colorado’s artificial intelligence law is being rewritten, and a new draft bill introduced on April 27, 2026, brings major changes for employers. Instead of focusing on whether a tool is “high-risk,” the proposal now targets how automated decisions affect people. This shift means employers must pay closer attention to hiring workflows, not just compliance paperwork.
Why Colorado Is Changing Its AI Law
Earlier this year, state lawmakers signaled a move away from the original Colorado Artificial Intelligence Act, which was set to take effect in June 2026. That law required heavy governance, including risk management programs and impact assessments for “high-risk” AI systems.
Now, a draft by Senator Rodriguez builds on earlier proposals with a more practical approach. The focus is no longer on labeling systems. Instead, it asks: Does the technology materially influence a big decision about someone’s life?
From System Rules to Decision Accountability
The old law treated AI as a category. The new draft treats it as a tool in a decision-making process. For example, if an automated system helps decide who gets a job, a loan, or housing, it falls under the new rules—even if it’s not traditional AI.
This shift reduces the need for formal AI governance programs. But it doesn’t remove compliance. Employers must now focus on how decisions are made, what role automated tools play, and how those decisions are explained to candidates.
Key Changes in the April 27 Draft
Here’s what the new draft does differently:
- No more risk management programs – The old requirement for detailed AI governance is gone.
- Notice required – Employers must tell candidates when automated tools are used in important decisions.
- Explanation for bad outcomes – If a tool leads to a negative decision (like not hiring someone), employers must explain its role.
- Right to correction – Individuals can ask to fix wrong data used in decisions.
- Human review – Candidates can request a meaningful review by a person.
These rules apply to decisions about employment, housing, credit, insurance, healthcare, and government services.
Who Is Responsible? A Clearer Liability Framework
The new draft also clarifies who is liable when things go wrong. It ties responsibility to how a system is designed, configured, and used. Companies can’t simply shift all blame to their vendors through contracts.
For employers using third-party hiring tools (like background check or scoring systems), this is a big deal. You need to know exactly how those tools work and what data they use. You can’t just rely on a vendor’s promises.
Existing Laws Still Apply
The bill doesn’t replace anti-discrimination or consumer protection laws. Employers must still follow rules like the Fair Credit Reporting Act (FCRA) for background checks. The new AI rules add extra layers of notice and explanation, but they don’t erase older obligations.
Where Things Are Still Unclear
Even with a more detailed draft, some questions remain:
- What does “materially influences” mean? – This is the trigger for the rules. In hiring, many inputs (interviews, background checks, skills tests) combine to form a decision. It’s not always clear when a tool crosses the line from “informing” to “influencing.”
- How does this work with FCRA? – Employers already send pre-adverse and adverse action notices for background checks. The new draft adds more disclosure requirements, but it’s not clear how to combine them without confusing candidates.
- Third-party data challenges – If you rely on a vendor for scoring or data, you may not have full control over how they explain their role. This could make compliance harder.
What Employers Should Do Now
The direction is clear: Colorado is moving toward decision-level accountability. Here are practical steps to prepare:
- Audit your hiring workflow – Identify every automated tool you use, from resume screeners to interview schedulers.
- Map decision points – Understand how each tool influences outcomes. Does it rank candidates? Flag concerns? Automatically reject?
- Review vendor contracts – Make sure your agreements clearly state who is responsible for disclosures and explanations.
- Update your notice process – Plan how you will inform candidates about automated tools and how to request human review.
- Train your team – HR and hiring managers need to understand the new rules and how to provide meaningful human review.
The Bottom Line
Colorado’s AI law is still evolving, but the direction is set. The focus is shifting from technology labels to real-world impact. For employers, compliance will live less in policy documents and more in how you structure hiring decisions. The question is no longer whether a tool is AI—it’s whether it shapes the outcome.
Stay tuned for further rulemaking and refinements. But don’t wait. Start reviewing your processes now to stay ahead of the 2026 timeline.
Comment