Colorado's new AI law reaches HOA screening decisions from January 2027
Colorado's new AI law reaches HOA screening decisions from January 2027
2026-09-10 · Colorado · Compliance
Colorado repealed and replaced its artificial intelligence statute in 2026, and the replacement has no small-entity exemption — which means an association or manager that uses a scoring tool to screen tenant applications or purchaser approvals is a regulated “deployer” from January 1, 2027. SB26-189 was signed by Governor Jared Polis on May 14, 2026. Its substantive duties take effect January 1, 2027 and apply to consequential decisions made on or after that date.1
Why an HOA is in scope at all
The act creates a new part 17 of article 1 of title 6, C.R.S. 6-1-1701 through 6-1-1709. Two definitions put associations inside it.
“Covered domain” at C.R.S. 6-1-1701(6) includes, at paragraph (c), “the lease or purchase of residential real estate in Colorado,” and at paragraph (b) “employment or an employment opportunity.” A “consequential decision” is one about a consumer relating to the provision of, or the consumer's access to, eligibility for, selection for, or compensation for a covered domain.
And a “deployer” at 6-1-1701(7) is simply “a person doing business in Colorado that deploys a covered ADMT.” There is no employee-count threshold and no small-entity carve-out anywhere in the part.
So an association — or a management company acting for it — that uses an automated tool to screen tenant applications under a leasing covenant, to score or rank a purchaser or transfer approval, or to screen applicants for an on-site job, is a deployer making consequential decisions, once the tool materially influences the outcome.
The carve-outs that will save most associations
“ADMT” excludes, at 6-1-1701(2)(b), calculators, databases, data storage, and “spreadsheets that require human analysis and do not use machine learning, foundation models, or large language models.” It also excludes a tool used by an individual “solely to summarize, organize, translate, draft, route, or present information for human review.” And “materially influence” at 6-1-1701(13) requires a non-de-minimis factor that constrains, ranks, scores, recommends, classifies or otherwise meaningfully alters how the decision is made — it “does not include incidental, trivial, or clerical uses.”
What a deployer actually owes
C.R.S. 6-1-1704 sets three obligations, and the second is the one that requires a template nobody has written yet.
A pre-decision notice. Before using a covered ADMT to materially influence a consequential decision, the deployer must give the consumer a clear and conspicuous notice that it did or will do so, plus how to get more information. Subsection (2) allows compliance through “a prominent public notice that is reasonably accessible at points of consumer interaction” — in practice, a line on the application form or the resident portal page.
A post-adverse-outcome disclosure within thirty days. Subsection (3) requires a plain-language description of the decision and the tool's role in it; instructions and “a simple-to-follow process” to request more information, including the tool's name, version number, developer, and “the types, categories, and sources of personal data used”; and an explanation of the consumer's rights under C.R.S. 6-1-1705.
Accessible delivery. Subsection (8) requires notices to be provided in a manner reasonably accessible to consumers with disabilities and consumers with limited English proficiency.
Two further duties sit outside 6-1-1704. C.R.S. 6-1-1703 requires records to be retained “for not less than three years after the date of a consequential decision” — longer than many associations' current retention schedules. And C.R.S. 6-1-1705 gives the consumer, after an adverse outcome, a right to correction of incorrect data and to meaningful human review. “Meaningful human review” at 6-1-1701(15) is defined: a designated individual with authority to approve, modify or override, who considers relevant available primary evidence, who is trained to conduct the review, and who “does not default to the system output.”
How it is enforced — and the trap in the enforcement clause
A violation is a deceptive trade practice under the Colorado Consumer Protection Act, enforceable exclusively by the Attorney General, with a right to cure. C.R.S. 6-1-1706 says plainly that “nothing in this part 17 creates a new private right of action.”
That is narrower relief than it looks. C.R.S. 6-1-1707(1) says a deployer “may be held liable in an action alleging unlawful discrimination under state anti-discrimination laws, including the ‘Colorado Anti-Discrimination Act’” arising from a consequential decision materially influenced by a covered ADMT. Subsection (9) closes the door harder: use of an ADMT “does not excuse, justify, or provide a defense to” any obligation or liability under state or federal law. And subsection (7)(a) voids as against public policy any indemnification clause that shifts a party's own anti-discrimination liability for ADMT use — so a management agreement that pushes that exposure onto the association is unenforceable as to that term.
What boards face before January 1, 2027
Inventory the AI in the stack, in writing. Ask every vendor — screening, collections scoring, violation detection or photo recognition, application triage, the resident portal — one question: does this product produce a score, ranking, recommendation, classification or prediction that we use in a decision about a lease, a purchase or transfer approval, or a job? The written answer determines whether the association is a deployer.
If the answer is yes, three deliverables. A pre-decision notice on the application or portal; a thirty-day adverse-outcome letter template; and a human-review path staffed by a named, trained person with override authority who does not rubber-stamp the model.
Set a three-year retention schedule for decision records, and push the developer disclosures upstream — C.R.S. 6-1-1702 obliges developers to give deployers the documentation a deployer needs, so put that obligation in the management agreement and the vendor contract.
The cheapest compliance route is to keep the human in front. If the tool only summarises or presents information for human review and emits no score, ranking or inference, it is outside the definition of ADMT altogether. For most Colorado associations, that is both the correct answer and the free one.
What to watch next
The Attorney General must adopt rules on or before January 1, 2027 clarifying the post-adverse-outcome disclosures under 6-1-1704(4)(b), and may adopt rules clarifying “materially influence” under 6-1-1706(5)(b). Those rules, not the statute, will determine how heavy this is for a volunteer board that has never thought of itself as an AI deployer.
Related Colorado HOA Topics
- SB26-189, "Automated Decision-Making Technology" — bill page, Colorado General Assembly ↩
- SB26-189 final act text, new C.R.S. 6-1-1701 through 6-1-1709 (source of the quoted definitions and duties) ↩
- 2026 Digest of Bills, Office of Legislative Legal Services (split effective dates) ↩
- Colorado General Assembly 2026 Legislative Updates, Colorado Division of Real Estate ↩
Stay on top of Colorado HOA law
Every week: new Colorado legislation, court rulings, and regulatory developments affecting condos, planned communities, and property managers. Free.
No spam. Unsubscribe anytime.