LR LossReserves.com
Subscribe
Back to The Wire Workers Comp

Colorado AI Law Rewrite Delays WC Claims Overhaul to 2027

SB 26-189 replaces Colorado's original AI anti-discrimination law with a lighter notice-based framework effective January 1, 2027, giving self-insured employers more time to evaluate how AI governance will reshape claims-handling severity savings.

By Sam, Editor May 2026 · 3 min read

Colorado Governor Jared Polis signed SB 26-189 on May 14, 2026, repealing the state’s original AI anti-discrimination law (SB 24-205) and replacing it with a narrower, notice-based framework that takes effect January 1, 2027. The replacement came after a federal magistrate judge blocked enforcement of the original law on April 27, following a constitutional challenge by xAI with Department of Justice support.

For self-insured employers whose TPAs deploy AI-driven claims tools on Colorado-sited workers’ comp claims, the rewrite buys time but does not eliminate the governance question.

What Changed

SB 24-205, signed in May 2024, would have required annual impact assessments, documented risk management programs, and human-in-the-loop oversight for every AI-influenced insurance decision. Those mandates carried real operational weight for WC claims operations: bill review algorithms, predictive case reserving models, and treatment authorization workflows all qualified as “high-risk” AI systems under the original law.

SB 26-189 strips out the three heaviest compliance burdens:

  • No more impact assessments. The annual documentation cycle that would have forced TPAs to audit every AI claims tool is gone.
  • No risk management program mandate. The original law’s requirement for a formal governance framework no longer applies.
  • No duty to prevent algorithmic discrimination. The “reasonable care” standard that created the largest single source of liability exposure has been removed.

What remains is a disclosure regime. Before using automated decision-making technology (ADMT) to materially influence a “consequential decision,” which includes insurance claims determinations, deployers must notify the consumer. Within 30 days of an adverse outcome, they must provide a plain-language explanation of the technology’s role. The Colorado Attorney General holds exclusive enforcement authority, and alleged violators get a 60-day cure window before any action proceeds.

The Insurance Safe Harbor

Insurers and TPAs already complying with Colorado’s existing algorithmic discrimination rules under Section 10-3-1104.9 are deemed compliant with SB 189’s requirements. Colorado’s Division of Insurance has had algorithmic governance rules in place since before SB 205 was written. Programs that already meet those baseline requirements face minimal incremental burden when the new law takes effect in January.

Who It Affects

Self-insured employers with Colorado-sited WC claims, particularly those in construction, energy, healthcare, and public entity sectors concentrated in the state. Any employer whose TPA uses AI-driven bill review, predictive triage, or automated case reserving on Colorado claims should confirm whether the TPA’s current compliance posture satisfies the safe harbor.

How It Hits Reserves

The original concern was that SB 205’s heavy governance mandates would slow AI processing and erode the medical severity savings that algorithmic bill review and claims adjudication have been delivering across WC programs. With SB 189’s lighter framework, those severity savings face less immediate regulatory drag in Colorado.

But the broader trend is not slowing. Twenty-four states have adopted the NAIC AI model bulletin, and several, including Florida’s HB 527, already prohibit algorithms as the sole basis for denying or reducing a claim payment. A patchwork of state-level AI governance rules is building regardless of Colorado’s course correction.

The reserve mechanism to watch is medical severity trend. If AI governance mandates eventually constrain algorithmic claims decisioning across multiple states, the development factors derived from the current AI-assisted processing environment may overstate the severity improvement embedded in recent accident years. Self-insured employers should ask whether their actuary’s medical severity trend picks assume continued AI-driven efficiency gains, and what happens to the IBNR estimate if those gains plateau.

What This Means for Your Next Review

Use the seven-month window before SB 189 takes effect to audit your TPA’s AI governance posture on Colorado claims. Confirm whether the safe harbor under existing algorithmic discrimination rules applies to your program. Ask your actuary whether medical severity trend assumptions incorporate AI-driven efficiency gains and whether those gains should be stress-tested against a scenario where multi-state AI regulation constrains claims automation.

Sources

More from The Wire
Workers Comp

Nevada's 21.6% WC Hike Exposes the Only Payroll Cap in NCCI

Nevada's largest-in-cycle WC loss cost increase took effect March 1, 2026, but the headline masks a structural payroll cap distortion that SB 317 will partially correct on October 1, creating a two-phase cost reset for self-insured employers.

3 min read Jun 2, 2026
Workers Comp

WCRI: WC Physician Prices 188% Higher in No-Fee-Schedule States

The 17th edition of WCRI's Medical Price Index quantifies an 18-year widening of the physician price gap between fee-schedule and non-fee-schedule states, with direct implications for medical severity assumptions in multi-state WC reserve studies.

3 min read Jun 2, 2026
Workers Comp

NCCI: WC Job Growth Quadruples, Signaling Exposure Rebound

NCCI's May 2026 labor market report shows employment growth accelerating from 10,000 to 76,000 jobs per month, with hiring concentrated in high-hazard WC sectors that drive payroll exposure bases higher.

3 min read Jun 1, 2026