30-YR FIXED6.71% +0.0515-YR FIXED6.04% +0.0610-YR TREASURY4.79% +0.0430-YR TREASURY5.27% +0.025-YR TREASURY4.55% +0.062-YR TREASURY4.39% +0.05FED FUNDS3.75% 0.00SOFR3.65% -0.01DOW53,062 +295S&P 5007,667 +35Freddie Mac · U.S. Treasury · Federal Reserve via FRED®30-YR FIXED6.71% +0.0515-YR FIXED6.04% +0.0610-YR TREASURY4.79% +0.0430-YR TREASURY5.27% +0.025-YR TREASURY4.55% +0.062-YR TREASURY4.39% +0.05FED FUNDS3.75% 0.00SOFR3.65% -0.01DOW53,062 +295S&P 5007,667 +35Freddie Mac · U.S. Treasury · Federal Reserve via FRED®
Thursday, September 3, 2026Bay Area Market: Coverage updated daily

Colorado’s AI proposal raises new compliance questions for lenders

MBA says definitions of ADMT and consequential decisions need clearer guidance for creditors

East Bay residential neighborhood, California
Curated News BriefBased on original reporting by HousingWire (August 25, 2026). The summary below is the Journal’s; the local analysis is original commentary by Omar Murillo.

According to HousingWire, Colorado has proposed new rules for automated decision-making technology that would apply to lending and other industries. The legislation, set to take effect on January 1, 2027, would require companies using this technology to follow stricter documentation, disclosure, and human review practices when those systems influence significant decisions about consumers, including mortgage lending.

The proposed rules define automated decision-making technology broadly as any system processing personal data to help make decisions about individuals in consequential matters like housing, lending, employment, and insurance. However, industry experts including the Mortgage Bankers Association have raised concerns that the definitions are too vague. The MBA specifically noted that the rules don't clearly explain what constitutes automated decision-making technology or when a decision becomes consequential, leaving lenders uncertain about which of their processes would actually be covered.

Under the new rules, technology developers would need to provide deployers with detailed documentation about their systems, including intended uses, training data categories, known limitations, and proper usage instructions. Both developers and deployers would need to keep records demonstrating compliance for at least three years. Lenders would also face new consumer disclosure requirements, needing to give clear notice before using such technology and providing explanations within 30 days if a decision results in an adverse outcome for the borrower.

The proposal also grants consumers the right to request human review of decisions made by automated systems, though lenders could potentially deny these requests if doing so would be commercially unreasonable. Consumers could also ask for corrections to inaccurate personal data used in the decision. Legal experts have noted that compliance will likely require lenders to thoroughly vet their technology vendors and establish new record-keeping processes.

One key concern from industry representatives is whether Colorado's requirements simply duplicate protections already provided by existing federal laws like the Equal Credit Opportunity Act and Fair Credit Reporting Act. They worry this could create overlapping compliance burdens rather than streamlining requirements. On a positive note for lenders, the Colorado law does not create a private right of action, meaning individual borrowers cannot sue directly for violations, with enforcement instead handled by the state attorney general.

What I am seeing locally here in the Bay Area is that we're often ahead of other regions when it comes to new regulations, so it's worth paying attention to what Colorado is doing. If these kinds of rules gain traction across states, lenders and real estate professionals will need to adapt quickly. For now, this highlights how important it is for borrowers to understand what technology is involved in their loan decisions and to know they have the right to ask for a human review when something doesn't feel right.