Colorado AI Regulation Delayed to June 30, 2026
Colorado Governor Polis signed SB 189 on May 14, 2026, which revises Colorado's original artificial intelligence law and delays the effective date from June 30, 2026, to January 1, 2027, while significantly scaling back its original requirements.
Objective Facts
On May 14, 2026, Colorado Governor Polis signed SB 189, which revises Colorado's original artificial intelligence law and delays the effective date from June 30, 2026, to January 1, 2027, while significantly scaling back its original requirements. The original Colorado AI Act, enacted in 2024, established a risk-based framework governing the use of AI in consequential decisions affecting areas such as employment, housing, health care and education. The new Act moves away from the original risk-based framework, eliminating the duty of care aimed at preventing algorithmic discrimination, deployer obligations to maintain risk management programs and conduct impact assessments, and certain reporting obligations to the Colorado Attorney General, adopting instead a narrower approach focused on disclosures and transparency around certain automated decision-making technologies. Senate Bill 189 passed the House by a 57-6 vote and cleared the Senate 34-1, with Senate Majority Leader Robert Rodriguez stating 'Everybody lost and everybody won.' The bill follows months of legal and political pressure against Colorado's original AI law, including President Donald Trump's December 2025 executive order targeting SB 205 as 'excessive State regulation' and a lawsuit filed by xAI challenging SB 205's constitutionality in April 2026.
Left-Leaning Perspective
Colorado AFL-CIO's Robert Lindgren spoke most forcefully against SB 189, arguing that 'We are talking about corporations with more wealth than most nations on Earth' and that the bill 'strips away rights' by eliminating 'risk management requirements, impact assessments, annual reviews and discrimination reporting' while introducing 'a cure period that lets developers delay accountability and allows discriminatory practices to continue under current law.' University of Colorado Boulder professor Jed Brown testified that he expected 'legions of complaints' from residents experiencing algorithmic discrimination and urged lawmakers to add a private right of action, stating 'The AG's office is not resourced for the massive scale of violations needing remedies.' Democratic Rep. Brianna Titone, the original 2024 bill's sponsor, explained that worker and consumer advocates faced pressure from Governor Polis, who did not want to sign without tech industry support, and acknowledged the final product 'wasn't the ideal bill advocates had hoped for, but has some aspects for consumers, like access to a human review, and more transparency.' The People's Alliance for Responsible Technology consortium, led by Kjersten Forseth of the AFL-CIO, adopted a more diplomatic public position, calling SB 189 'a good first step to protect the interests of everyday Coloradans from some of the negative consequences of AI' while noting 'there's still work to be done to ensure Big Tech companies are transparent and accountable' and expressing encouragement about 'a path to hold developers accountable when their technologies result in discrimination and other harm.' Left-leaning critics of SB 189 emphasize the removal of protective mechanisms: the elimination of mandatory bias audits, impact assessments, and the duty of reasonable care leave algorithmic discrimination unchecked. The emphasis on enforcement through Attorney General discretion rather than private lawsuits, combined with the temporary cure period, represents a structural weakening that labor groups argue disproportionately favors large technology companies over workers and vulnerable consumers.
Right-Leaning Perspective
Brittany Morris Saunders, CEO of the Colorado Technology Association and a central figure in SB 189 negotiations, hailed the bill as 'meaningful progress for Colorado and a more balanced path forward,' emphasizing that 'CTA has worked with policymakers, member companies, and partners across the business and technology communities to advance a framework that protects consumers while allowing Colorado companies to innovate, hire and grow.' Bryan Leach, CEO of Ibotta, characterized SB 189 as 'a marked improvement over the original bill' and 'a much more practical approach,' though he expressed concerns about specific provisions like the sunset of the right-to-cure clause. The Colorado Chamber of Commerce President stated the bill 'found the balance that addresses the interests of businesses and consumers and keeps Colorado a place where companies want to do business.' The right-leaning case is reinforced by the original law's reception: the tech industry consistently argued the original law was unworkable and would stifle innovation; President Trump's December 2025 executive order threatened to withhold federal funds from Colorado if it enforced SB 205, calling it 'cumbersome' in his directive to bar states from 'robust AI regulations'; and Elon Musk's xAI company filed suit challenging SB 205's constitutionality. Business and technology leaders, along with Governor Polis himself, criticized the 2024 law as being too burdensome for AI developers and deployers, requiring both parties to complete extensive risk analyses to determine the possibility of algorithmic discrimination. Right-leaning voices stress practical workability: the original law's comprehensive risk assessment requirements, impact assessments, and duty-of-care provisions created compliance burdens that threatened innovation and company operations. SB 189's shift to a notice-and-transparency model allows the AI industry to flourish while retaining baseline consumer protections.
Deep Dive
Colorado's AI regulation saga reflects a clash between two competing policy visions that proved impossible to reconcile in a single statute. The original 2024 Colorado AI Act represented the nation's most comprehensive attempt to prevent algorithmic discrimination through proactive governance—mandatory risk assessments, impact audits, and a duty of reasonable care. This approach drew international comparison to the EU AI Act's risk-based framework. However, from the moment Governor Polis signed it, he signaled dissatisfaction. Business leaders warned the requirements were technically unworkable and threatened to relocate operations. The Trump administration's December 2025 executive order, combined with xAI's constitutional challenge, created external pressure that matched internal industry lobbying. What each side gets right: Labor and consumer advocates correctly identify that shifting from preventative safeguards to post-hoc remedies reduces the barrier to entry for discriminatory systems. Studies cited by academics like Jed Brown do show widespread algorithmic bias in hiring, lending, and housing decisions. The AFT-CIO's point that enforcement depends entirely on an under-resourced Attorney General's office is substantive—Colorado's AG has no dedicated AI enforcement staff. Technology industry advocates, meanwhile, accurately note that the 2024 law's dual requirements (impact assessments plus reasonable care duty) created overlapping compliance burdens that no AI deployment system had been designed to accommodate. The practical difficulty of proving 'algorithmic discrimination risk' in advance is real; SB 189's shift to transparency and after-the-fact disclosure is operationally simpler. What both sides omit: Left-leaning critics underestimate the genuine consensus-building work that occurred in Polis's private working group—this was not a unilateral industry capture but a multi-month negotiation where unions and consumer groups signed off on the framework. Right-leaning voices downplay that the original law's deterrent effect (risk of attorney general action) is eliminated; without proactive impact assessment requirements, many systems will proceed to deployment with known bias risks simply flagged in fine-print disclosures. What happens next matters enormously. The Attorney General must adopt rules clarifying post-adverse-outcome disclosure requirements and consumer rights by January 1, 2027. The xAI litigation continues; if courts rule the original law unconstitutional on vagueness grounds, SB 189 could face the same challenge. Colorado's fate signals to other states whether comprehensive AI regulation is politically survivable or will collapse under industry pressure. The broad bipartisan votes masking deep disagreement suggest the 'consensus' was acceptance of exhaustion rather than genuine agreement. Labor's tepid public support ('good first step') reflects strategic calculation that some oversight beats none. How Colorado enforces SB 189 over the next 18 months will determine whether it becomes a template for other states or a cautionary tale of regulatory capture.