Maryland insurance AI regulation

Maryland has adopted the NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, via Bulletin No. 24-11, adopted 2024-04-22.

Details

Jurisdiction	Maryland (MD)
NAIC model bulletin	Adopted
Citation	Bulletin No. 24-11
Date	2024-04-22 (adopted)
Source	Implementation of NAIC Model Bulletin: Use of AI Systems by Insurers, as of 2026-08-06

Other AI-related insurance activity

Measure	Status
HB 1563 (2026 Regular Session), Chapter 165 QUARTERLY carrier reporting to the Insurance Commissioner. The report must include the number of adverse decisions issued under § 15-10A-02(f), whether the adverse decision involved a prior authorization or step therapy protocol, the type of service at issue, and whether an artificial intelligence, algorithm or other software tool was used in making the adverse decision. (a)(1) On a quarterly basis, each carrier shall submit to the Commissioner ... the number of adverse decisions issued by the carrier under § 15-10A-02(f) of this subtitle, whether the adverse decision involved a prior authorization or step therapy protocol, the type of service at issue in the adverse decisions, and whether an artificial intelligence, algorithm, or other software tool was used in making the adverse decision source	enacted, effective 2026-06-01 The act's title concerns emergency room and post-acute care coverage and gives no indication of an AI provision. The duty is real and is easy to miss from the title alone — which is how it came to be disputed here. ✓ verified against the source text
HB 820 (2025 Regular Session), Chapter 747 Substantially more than a guardrail adoption. Carriers, pharmacy benefits managers and private review agents using AI, an algorithm or other software tool for utilization review must base determinations on the individual patient's medical history, clinical circumstances and relevant records, and may NOT base decisions solely on group datasets. AI may not replace provider decision-making, must not discriminate unfairly against enrollees, and must be applied equitably. Carriers must maintain WRITTEN POLICIES for AI use, conduct QUARTERLY REVIEWS of AI performance, protect patient data privacy, and ensure the tools do not cause patient harm. AI may not deny, delay or modify health care services. Correction: Previously recorded only as 'adopted the CMS guardrails'. That understated it severely: this is a dedicated AI utilization-review statute with written-policy and quarterly-review duties. Together with HB 1563's quarterly reporting, Maryland has the most developed AI utilization review regime of any state tracked here. source	enacted, effective 2025-10-01 ✓ verified against the source text

Measure

Status

HB 1563 (2026 Regular Session), Chapter 165
QUARTERLY carrier reporting to the Insurance Commissioner. The report must include the number of adverse decisions issued under § 15-10A-02(f), whether the adverse decision involved a prior authorization or step therapy protocol, the type of service at issue, and whether an artificial intelligence, algorithm or other software tool was used in making the adverse decision.

(a)(1) On a quarterly basis, each carrier shall submit to the Commissioner ... the number of adverse decisions issued by the carrier under § 15-10A-02(f) of this subtitle, whether the adverse decision involved a prior authorization or step therapy protocol, the type of service at issue in the adverse decisions, and whether an artificial intelligence, algorithm, or other software tool was used in making the adverse decision

source

enacted, effective 2026-06-01
The act's title concerns emergency room and post-acute care coverage and gives no indication of an AI provision. The duty is real and is easy to miss from the title alone — which is how it came to be disputed here.
✓ verified against the source text

HB 820 (2025 Regular Session), Chapter 747
Substantially more than a guardrail adoption. Carriers, pharmacy benefits managers and private review agents using AI, an algorithm or other software tool for utilization review must base determinations on the individual patient's medical history, clinical circumstances and relevant records, and may NOT base decisions solely on group datasets. AI may not replace provider decision-making, must not discriminate unfairly against enrollees, and must be applied equitably. Carriers must maintain WRITTEN POLICIES for AI use, conduct QUARTERLY REVIEWS of AI performance, protect patient data privacy, and ensure the tools do not cause patient harm. AI may not deny, delay or modify health care services.
Correction: Previously recorded only as 'adopted the CMS guardrails'. That understated it severely: this is a dedicated AI utilization-review statute with written-policy and quarterly-review duties. Together with HB 1563's quarterly reporting, Maryland has the most developed AI utilization review regime of any state tracked here.
source

enacted, effective 2025-10-01
✓ verified against the source text

Context

Of 51 tracked US jurisdictions, 25 have adopted the NAIC model bulletin, 4 run their own insurance-specific AI framework (California, Colorado, New York and Texas), and 22 have not adopted it but have other tracked AI-related insurance activity. The bulletin restates that existing unfair trade practice and unfair discrimination law applies to AI-driven decisions; it does not create new statutory obligations.

Two tracks. Insurance AI regulation runs on two largely independent tracks: the NAIC model bulletin (governance, all lines), and a 2026 wave of health-insurance statutes restricting AI as the sole basis for claim or coverage denial. A state can be active on the second while absent from the first. Absence from the NAIC map is not absence of regulation.

See all 51 jurisdictions · This record as JSON