What is the Colorado AI disclosure law? The Colorado AI disclosure law, formally SB 26-189, is the rewritten 2024 Colorado AI Act passed by the General Assembly on May 12, 2026, requiring any operator, including commercial real estate lenders, landlords, and property managers, to notify consumers whenever automated decision-making technology materially influences a consequential decision about housing, financial or lending services, employment, or insurance. The law pushes the original June 2026 effective date to January 2027 and gives consumers the right to know an AI was used, see what personal data fed the decision, correct inaccuracies, and request a human re-review. For a full landscape view of how this fits the broader regulatory shift, see our complete guide on the best AI tools for commercial real estate investors.
Key Takeaways
- Colorado SB 26-189 replaces the 2024 Colorado AI Act with a narrower disclosure regime focused on automated decision-making technology that materially influences consequential decisions, including housing and lending.
- The new effective date is January 2027, pushed back from the original June 2026 deadline. CRE operators have roughly twenty months to build disclosure, appeals, and data-correction workflows.
- House passed SB 26-189 by 57 to 6 and the Senate by 34 to 1. Governor Jared Polis said he will sign it. Colorado Chamber of Commerce CEO Loren Furman called it a model other states will copy.
- Any AI tool used to score tenants, underwrite loans, value collateral, or recommend renewal pricing for Colorado properties triggers disclosure if the output materially influences a decision affecting a Colorado resident.
- The law dropped developer training-data disclosures and testing requirements present in the original act, lowering implementation cost but raising appeal volume risk for property managers and lenders.
What SB 26-189 Actually Changes
The original 2024 Colorado AI Act, SB24-205, would have forced developers and deployers to disclose how high-risk AI systems were trained, document algorithmic impact assessments, and publish public risk summaries. That framework drew opposition from xAI, which filed suit in April 2026 to block enforcement, and from the Trump Administration, which threatened to withhold federal funds from states with aggressive AI rules. SB 26-189, sponsored by Senate Majority Leader Robert Rodriguez, scraps the high-risk system category and replaces it with a simpler concept: automated decision-making technology, or ADMT, that processes personal data used to materially influence a consequential decision. Consequential decisions are defined as those affecting access to, eligibility for, or compensation related to education, employment, housing, financial or lending services, insurance, health care, or essential government services. Operators must notify consumers when ADMT is used, let them see the personal data considered, correct inaccuracies, and request a human re-review.
Why CRE Investors and Operators Should Care
Most commercial real estate workflows now run on or alongside AI. CBRE, JLL, and Cushman & Wakefield publish AI-driven market intelligence used in CRE underwriting. AppFolio, Yardi, RealPage, and Buildium ship AI-powered tenant screening, lease abstraction, and renewal pricing modules. Lenders use ChatGPT and Claude to draft credit memos, score borrowers, and analyze rent rolls. Each of those touch points becomes a regulated activity in Colorado the moment the output materially influences a decision about a Colorado resident's housing or credit. According to Gallup's Q1 2026 Workforce survey, AI use in the workplace has reached 50%, up from 21% in Q2 2023, which means even informal use of ChatGPT to screen a Colorado tenant application can fall inside the law if the output drives the decision. The AI in real estate market is projected to reach $1.3 trillion by 2030, and 92% of corporate occupiers have already initiated AI programs, per JLL research. Colorado is now the first state to attach individual consumer rights directly to those workflows.
Three CRE Workflows That Now Require Disclosure
- AI Tenant Screening: If you use AppFolio, RealPage, Findigs, TurboTenant, or a custom ChatGPT prompt to score rental applicants for a Colorado property, you must notify the applicant that ADMT influenced the decision. See our commercial tenant screening guide for workflow examples.
- AI Loan Underwriting and Credit Decisions: Lenders using Claude or Gemini to analyze DSCR, debt yield, or borrower credit for a Colorado property must disclose AI involvement, preserve the underlying data, and stand up an appeals path. Cash-on-cash and DSCR calculations driven by AI agents fall squarely inside the consequential-decision definition.
- AI Valuation and Property Condition Scoring: Tools like Restb.ai computer vision, CubiCasa floor plans, and AI-assisted broker opinions of value used to set a price that affects a homeowner or seller in Colorado trigger disclosure when the value materially shapes a financing or sale decision.
Implementation Checklist for CRE Operators
Start work now. Tenant and borrower volume in any active CRE portfolio means thousands of notice events in the first quarter of 2027 alone. A practical checklist:
- Inventory every AI touchpoint: Map each AI tool that touches a Colorado resident decision. ChatGPT, Claude, Gemini, Perplexity, Yardi RentMaximizer, RealPage AI Revenue Management, and AppFolio AI Leasing all count.
- Draft a plain-language ADMT notice: One paragraph at the top of every adverse-action letter, lease decline, or loan denial. Include what AI was used, what data fed it, and how to request human re-review.
- Build a thirty-day appeals workflow: Route appeals to a named human reviewer with authority to overturn the AI output. Track resolution times.
- Preserve the data trail: Store the inputs the AI saw, the prompt or rule set, and the output. Six-year retention aligns with most CRE recordkeeping standards.
- Update vendor contracts: Require AppFolio, Yardi, RealPage, and any other AI vendor to indemnify you for ADMT compliance and provide the data-access functionality consumers can invoke.
For personalized guidance on implementing these workflows in your portfolio, connect with The AI Consulting Network.
How Colorado Compares to Connecticut SB 5
Connecticut passed its own omnibus AI law, SB 5, on May 1, 2026, with staggered effective dates running October 2026 through October 2027. Connecticut SB 5 leans heavier on employment-related AI obligations, chatbot safety, and synthetic content labeling. Colorado SB 26-189 leans heavier on housing and financial-services decisions, with a tighter focus on individual consumer rights. CRE operators doing business in both states will need two disclosure templates and two appeals workflows. For the full breakdown of Connecticut's framework, see our analysis of the Connecticut SB 5 AI Responsibility Act. Expect more states to follow the Colorado disclosure template through 2026 and 2027, because the Colorado Chamber of Commerce, Senator Rodriguez, and the Trump Administration all signaled it is the politically viable middle ground.
What This Means for AI Tool Selection in 2026
Operators evaluating new AI tools should now ask three questions before signing a vendor contract. First, does the tool log inputs and outputs in a format you can hand a consumer on demand. Second, does it expose the variables that drove the decision in plain language. Third, does the vendor contractually agree to keep that data accessible for the full Colorado retention window. Tools that fail any of the three create disclosure debt that compounds with every Colorado tenant or borrower interaction. CRE investors looking for hands-on AI implementation support can reach out to Avi Hacker, J.D. at The AI Consulting Network.
Frequently Asked Questions
Q: Does Colorado SB 26-189 apply to small landlords with only a few units?
A: Yes, if the landlord uses AI or automated decision-making technology to make a consequential decision about a Colorado resident. The law focuses on the type of decision, not the size of the operator. Small landlords using ChatGPT to screen tenants or set rents are inside the scope.
Q: When does Colorado SB 26-189 take effect?
A: January 2027. The original Colorado AI Act was scheduled to take effect in June 2026, but SB 26-189 pushed the date back to January 2027 to give operators time to build disclosure and appeals workflows.
Q: What counts as a consequential decision under SB 26-189?
A: Any decision that materially influences a Colorado consumer's access to, eligibility for, or compensation related to housing, financial or lending services, employment, insurance, education, health care, or essential government services. For CRE, that means tenant approvals, loan decisions, rent pricing, and property-conditioned valuations all qualify.
Q: Do I have to disclose how my AI tool works?
A: No. SB 26-189 dropped the training-data and algorithmic-impact disclosures from the original 2024 law. You only have to notify consumers that AI was used, let them see the personal data considered, correct inaccuracies, and offer a human re-review. If you are ready to transform your underwriting and screening processes with compliant AI workflows, The AI Consulting Network specializes in exactly this.
Q: Will other states copy Colorado's approach?
A: Industry observers expect yes. Colorado Chamber of Commerce CEO Loren Furman publicly described SB 26-189 as a model for other states, and the disclosure-only framework avoids the federal preemption fight that derailed the original 2024 law. Connecticut SB 5, Iowa's chatbot law, and active bills in Utah and Vermont all signal that 2026 is the year state AI laws stop being theoretical for CRE operators.