Colorado Delays AI Law to June 2026 – Top 10 Questions for Employers and Tech Developers
Insights
8.27.25
Colorado lawmakers couldn’t reach a compromise to refine the nation’s first statewide AI antidiscrimination law and instead agreed yesterday to delay the law’s implementation date from February to June 2026 – all in hopes that next year’s legislative session will yield substantive changes before it takes effect. There were plenty of twists and turns during the brief special legislative session that just wrapped. For a brief time it appeared lawmakers were going to agree to a watered-down version that would have cut tech developers a big break but still saddled employers and businesses with plenty of obligations. But after negotiations broke down, the legislature decided to kick the can down the road and look to the 2026 legislative session as the time to work out potential revisions. What are the 10 biggest unanswered questions we have about Colorado’s groundbreaking law?
Quick Resource Center
|
What Happened This Week?
Over the weekend, it appeared that lawmakers agreed to a compromise measure that would have eased burdens on the tech community but retained many employer and business obligations.
- But the grand bargain fell apart on Monday, leading the state Senate to pass a measure (29-3) to instead delay the implementation date from February 1, 2026, to June 30, 2026.
- The state House passed the measure (48-14) on Tuesday.
- Governor Polis is expected to sign the delay into effect in short order.
Top 10 Unanswered Questions
While we now have some additional breathing room, a number of questions remain unanswered. Here are our top ten questions leading into the 2026 legislative session.
1. What will happen when the legislature reconvenes in 2026?
This is the million-dollar question. Just as we predicted at the beginning of this special session, we expect there to be three possible pathways when Colorado’s AI law is up for debate again in January:
- Lawmakers could kick the can down the road again and delay implementation for another period of time to allow further assessment of the state of technology and the need for strict regulation.
- The law could be refined, perhaps to look closer to the compromise measure that almost passed this week.
- We may see a consensus build that it’s time to just bite the bullet and allow the law to take effect on June 30, 2026.
2. What impact will other states’ activity have on Colorado?
By the time the Colorado legislature reconvenes in January 2026, we may be looking at another state leading the way when it comes to AI regulation: California. Lawmakers in Sacramento are currently debating several measures that could lead to strict AI regulation on tech developers, employers, and businesses. We’ll know by September 12 whether any of these will land on the desk of the Governor for action. If California passes similar legislation, that could lead Colorado lawmakers to want to keep pace – or pump the brakes and wait to see the effect it has on California businesses before charting their own course.
3. Will a revised measure ease obligations of tech developers?
Colorado lawmakers believed they had done enough to calm the concerns of the tech community with their compromise measure. The proposal would have removed responsibilities related to algorithmic discrimination, risk management programs, self-reporting, and more. But that wasn’t enough for tech leaders, who still balked at the disclosure obligations and liability concerns that remained in the compromise. How lawmakers approach this issue in 2026 will be a critical issue to watch for.
4. Will we still see joint and several liability between developers and deployers (employers and businesses)?
The compromise measure clarified that AI developers and deployers (employers and businesses using AI) would share “joint and several” liability for any violations of the AI law. That means that the state Attorney General could have pursued relief for alleged AI discrimination against applicants, workers, consumers, and others by bringing action against the developer of the AI system and the business that used it and found them equally liable. We’ll see whether this same system is included in the next iteration of the proposed law or if developers catch a break and gain some additional protections.
5. Will employers and businesses still need to proactively produce risk assessments and impact reports?
One of the positive developments for AI deployers (those employers and businesses using AI) in the compromise measure was the removal of the obligation to produce risk assessments and impact reports about the AI systems they were using. The business community will eagerly await 2026 to see if the next proposal also removes these burdensome obligations.
6. What will employers need to do when it comes to notifying applicants and workers about the use of AI?
We’ll have to wait until 2026 to see whether lawmakers still want them to notify applicants and workers about the use of AI during hiring and employment – and also offer them appeal and data correction rights.
7. Will businesses still need to disclose that AI is fueling website chatbots?
Both the existing law and the compromise measure required businesses to disclose to website/app users when a chatbot or other tool was actually run by artificial intelligence and not a human – even if it was obvious to a reasonable person that they were talking to AI. It stands reasonable to predict that a similar obligation will remain in a 2026 version of the law.
8. Will a revised law include a private right of action?
Neither the original law nor the compromise proposal allowed individuals to bring suit against AI developers, employers, or businesses using AI (also known as a “private right of action”). Instead, sole responsibility for enforcement would be held by the state Attorney General. It seems likely that any new version of the law will retain this same structure, but the business community will be waiting patiently for confirmation.
9. Will Colorado’s law become the de facto standard for the country?
Employers and businesses headquartered outside Colorado will almost certainly be affected by any new law passed in 2026. We expect it will apply to businesses making employment decisions about candidates or employees who reside in Colorado, even if the company is located elsewhere. So the question arises: will multi-state employers build a Colorado-specific compliance program, or implement uniform practices across their entire US workforce to avoid running parallel systems? If businesses choose the latter approach for operational efficiency, will Colorado’s law become a de facto national standard?
10. Will Congress step in and push states to scrap their AI laws?
Finally – will all of this discussion be rendered irrelevant by Congress? Federal lawmakers came very close to passing a law last month that would have either blocked states from developing their own AI-related laws or blocked states from receiving massive amounts of federal funding if they enforced AI-related laws. Speaking at FP’s AI Conference last month, Rep. Jay Obernolte (R-CA), the Co-Chair of the bipartisan House Task Force on Artificial Intelligence, seemed inclined to believe that such an “AI pause” could reemerge later this year or in 2026.
Conclusion
Make sure you are subscribed to Fisher Phillips’ Insight System to receive the most up-to-date information directly to your inbox. We will continue to monitor the situation and provide updates as they unfold. For more information, contact your Fisher Phillips attorney, the authors of this Insight, any attorney in our Denver office, or any attorney in our AI, Data, and Analytics Practice Group.
Related People
-
- Michael R. Greco
- Regional Managing Partner
-
- Richard R. Meneghello
- Chief Content Officer