Insurance: health insurers; use of artificial intelligence tools to make decisions regarding claims; prohibit. Amends 1956 PA 218 (MCL 500.100 - 500.8302) by adding sec. 3406ss.

2025-2026 Regular Session Introduced by Kelly Breen and 10 co-sponsors

HB 4536 bars Michigan insurers from denying, delaying, or modifying claims solely due to an AI-based review; requires safeguards and human oversight in adjudication.

re-referred to Committee on Communications and Technology

#communications:artificialintelligence #insurance:healthinsurers

Summary — HB 4536 (Insurance: prohibition on AI-based claim denials/modifications/delays)

Bill number: HB 4536
Sponsor (introduced version): Rep. Carrie Rheingans
Date introduced / filed: Filed March 12, 2025; House introduced version dated June 3, 2025
Placement/status: Referred to Committee on Insurance (most recent action Sept. 18, 2025). See timeline below.
Statutory change proposed: Adds section 3406ss to the Insurance Code of 1956 (MCL 500.100–500.8302).

Purpose / Intent

The bill seeks to prevent health insurers from taking adverse claim actions (denying, modifying, or delaying payment) when those actions are based on a review done using artificial intelligence (AI). It is intended to limit or govern the role of automated decision-making tools in health-claims adjudication.

What the bill would change / require

Prohibits health insurers operating in Michigan from denying, altering, or postponing payment on an insurance claim when the reason for that action is a review performed using AI.

The prohibition applies to policies delivered, issued for delivery, or renewed in Michigan.

The bill is narrow in scope: it targets claim decisions tied to an AI review rather than a broader ban on using AI in operations.

Who is affected

Health insurers doing business in Michigan (policy issuance, delivery, renewals).

Insured individuals and beneficiaries — protections against claims denials/modifications/delays based solely on AI reviews.

Providers and third‑party administrators who use insurer adjudication systems.

Vendors of AI claim‑review tools.

Notable omissions / implementation questions

The bill does not define "artificial intelligence" or specify what constitutes a "review using artificial intelligence."

It does not state enforcement mechanisms, penalties, or remedies for violations.

It does not clarify whether AI-informed decisions combined with human review are covered, or whether AI may be used for non‑adjudicative functions (e.g., fraud detection, process automation).

Regulatory guidance or future amendments would likely be needed to address definitions, compliance processes, and oversight.

Procedural timeline (selected actions)

2025-03-12: Filed

2025-04-03: Read first time; referred to State Affairs (initial calendar activity)

2025-04-23–24: Committee hearing; reported favorably without amendment

2025-04-29: Committee report sent to Calendars

2025-06-03: House introduced version by Rep. Rheingans (referred to Communications & Technology)

2025-09-18: Rule suspended; motion to discharge committee approved; placed on second reading; referred to Committee on Insurance (current referral)

Practical impact

If enacted as written, insurers would need to ensure that any adverse claim decisions are not based solely on AI reviews. This likely would require changes to adjudication workflows (e.g., human review safeguards or exclusion of AI from final decision-making). Because the bill lacks definitional and enforcement detail, additional legislative or regulatory action would probably follow to clarify how it is implemented and enforced.

Bill