AI in Utilization Review: Current Uses and Regulations

Artificial intelligence is becoming more visible across healthcare and claims operations, including utilization review.

As AI tools become more integrated into utilization review workflows, regulators are starting to define where technology can support the process, and where clinical decision-making must remain human-led. Understanding both sides of that conversation is essential.

How Utilization Review Works

Utilization review is the process of evaluating medical services against established clinical criteria to determine medical necessity. While requirements vary by state, UR is generally governed by state regulations, guided by evidence-based frameworks such as ODG or ACOEM, and supported by licensed clinical reviewers throughout the decision-making process. That structure matters when evaluating AI adoption because most current regulations are focused on how AI is used and where clinical responsibility remains.

How AI Is Being Used in Utilization Review Today

AI applications in utilization review generally fall into several categories, ranging from administrative support to tools that assist reviewers during clinical evaluation.

Document Organization and Information Extraction

One of the most common current applications is organizing incoming documentation.

AI tools can:

• Sort medical records
• Extract relevant clinical details
• Identify diagnosis and treatment codes
• Flag missing documentation
• Generate structured summaries for review teams

The result is less time spent navigating records and more time focused on clinical review.

Guideline Comparison and Request Flagging

AI can also compare treatment requests against established clinical guidelines such as ODG, ACOEM, or state-required protocols.

These tools may flag requests that appear to meet criteria, identify those that fall outside guidelines, and surface cases that require closer review. In practice, this serves more as a prioritization and workflow support function than a replacement for the review process itself.

Outlier Detection and Pattern Recognition in Utilization Review

Some AI tools are being applied across broader claims workflows with direct implications for utilization review.

These systems can identify treatment patterns that deviate from expected norms for a diagnosis or injury type, such as:

• Referral activity that trends significantly above expected ranges
• Treatment utilization that escalates without supporting documentation

The goal is to highlight cases that may warrant additional evaluation.

Predictive Triage in Utilization Review

Predictive analytics tools are increasingly being used to identify claims that may become high-cost or clinically complex earlier in the lifecycle.

The 2025–2026 Workers’ Compensation Industry Insights Survey by Risk & Insurance found that respondents identified fraud, waste, and abuse detection as the top short-term application for AI, followed by summarizing and sharing medical records.

Current Utilization Review Regulations

Regulatory attention around AI in utilization review has focused on maintaining human accountability in medical necessity decisions.

As regulations continue to emerge, the focus is still establishing where human accountability must stay in place regardless of what technology supports the process.

California SB 1120

As of early 2026, California SB 1120 is the only enacted legislation in the United States that directly regulates AI use in utilization review. The law requires that AI tools used in UR support — not replace — the judgment of a licensed physician or qualified healthcare professional. Final medical necessity determinations must be made by a qualified human clinician, and AI use must meet standards for fairness and equity. The law applies to health plans operating in California and is widely cited as a model for similar legislation in other states.

Other State Regulations Around Utilization Review and AI

Texas followed California’s lead, passing a law that regulates AI use in UR for health benefit plans. New York, Rhode Island, and Tennessee have each introduced similar legislation, with human accountability and transparency as the consistent themes across all of them.

NAIC Guidance for AI Oversight in Utilization Review

The National Association of Insurance Commissioners issued a Model Bulletin in December 2023 recommending that insurers maintain governance frameworks for AI tools, regularly test models for bias, and ensure human accountability for AI-influenced decisions.

The bulletin is not binding law, as state adoption remains voluntary, but it is the regulatory baseline most state insurance departments are referencing. Notably, the NAIC’s 2025 survey found that nearly one-third of health insurers still do not regularly test their AI models for bias, despite the bulletin’s guidance

Final Thoughts

AI is a legitimate and growing part of UR operations, but the regulatory direction is unambiguous. Two states have now codified into law what the NAIC has been recommending since 2023: a human clinician makes the call, and that accountability cannot be delegated to an algorithm.

For claims and medical management professionals, the question to ask about any AI tool in a UR workflow isn’t whether it improves efficiency; it’s whether the process still puts a qualified clinician in genuine decision-making authority.

Looking for UR solutions backed by clinical expertise? Connect with us today.

Check out our sources:

Bean, Melissa, et al. “Utilization Review in Workers’ Compensation: Review of Current Status and Recommendations for Future Improvement.” Journal of Occupational and Environmental Medicine, vol. 62, no. 6, June 2020, pp. e273–e286. American College of Occupational and Environmental Medicine, https://acoem.org/acoem/media/PDF-Library/Publications/Utilization_Review_in_Workers__Compensation_-15.pdf.

Becker, Josh. “SB 1120: Health Care Coverage: Utilization Review.” California Legislative Information, approved 28 Sept. 2024, https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB1120.

“Incorporating Artificial Intelligence (AI) in the Utilization Review (UR) Process.” National Association of Independent Review Organizations (NAIRO), 29 July 2025, https://www.nairo.org/index.php?option=com_dailyplanetblog&view=entry&year=2025&month=07&day=29&id=25:incorporating-ai-in-ur.

“Model Bulletin on the Use of Artificial Intelligence Systems by Insurers.” National Association of Insurance Commissioners (NAIC), adopted 4 Dec. 2023, https://content.naic.org/sites/default/files/inline-files/2023-12-4%20Model%20Bulletin_Adopted_0.pdf.

“NAIC Health Insurance AI/ML Survey Report.” National Association of Insurance Commissioners (NAIC), May 2025, https://content.naic.org/insurance-topics/artificial-intelligence.

“SB 815: Relating to the Use of Certain Automated Systems in, and Certain Adverse Determinations Made in Connection with, the Health Benefit Claims Process.” Texas Legislature Online, 89th Legislature, enrolled 20 June 2025, effective 1 Sept. 2025, https://capitol.texas.gov/BillLookup/History.aspx?LegSess=89R&Bill=SB815.

“Survey Reveals Workers’ Comp Industry Faces Rising Complexity, Shifting Technology Priorities.” Risk & Insurance, 15 Jan. 2026, https://riskandinsurance.com/survey-reveals-workers-comp-industry-faces-rising-complexity-shifting-technology-priorities.