Michigan AI Laws: What's Enacted and What's Next (2027)

As of November 2026, Michigan has enacted targeted AI laws but, in our search, no comprehensive AI act. The main laws we found are the 2023 election package (HB 5141, 5143, 5144 and 5145, Public Acts 263-266 of 2023) and the Protection from Intimate Deep Fakes Act (HB 4047, Public Act 11 of 2025).

Status as of November 4, 2026: The 2023 and 2025 laws are in force. SB 760, the chatbot bill for minors, passed the Senate 20-17 on April 29, 2026 and is in a House committee, so it is not law. HB 4668 (frontier AI safety), HB 5579 and SB 1077 (workplace AI), and HB 4536 and 4537 (AI in claims decisions) are in committee. HB 5899 (a state AI pilot program) was referred to the House Rules committee.

This page covers Michigan statutes and bills as of November 4, 2026, and one insurance regulator's release about a bulletin. It is general legal information, not legal advice. It does not survey governor's orders or directives, state agency policies, other regulator guidance, or Attorney General enforcement, so a question answered here as "no statute" may still be addressed by a regulator or other executive action. For the national picture, see our guide to US AI laws.

Michigan AI Laws at a Glance

Does Michigan Have a Comprehensive AI Law?

We found none. We also found no enacted Michigan statute written specifically for frontier AI models, AI in hiring, AI in health care, or companion chatbots. Other enacted provisions we found that mention AI fund education or training work and do not set rules for private conduct.

Those are results of one search, not legal conclusions. We searched the full text of the Michigan Compiled Laws, complete through Public Act 114 of 2026, and the Legislature's public act and bill search, for terms including "artificial intelligence," "chatbot" and "automated decision." A law that uses other words would not surface in that search.

Michigan's AI Political Ad Disclosure Law (HB 5141, MCL 169.259)

The Governor approved the four-bill election package on November 30, 2023, and it took effect February 13, 2024. Public Act 263 added MCL 169.259 to the Michigan Campaign Finance Act.

The duty falls on a person, committee or other entity that "creates, originally publishes, or originally distributes" a qualified political advertisement. The ad must state, clearly and conspicuously, that it "was generated in whole or substantially by artificial intelligence."

A "qualified political advertisement" is limited. It means a paid advertisement relating to a candidate for federal, state or local office in Michigan, such an election, or a ballot question, "that contains any image, audio, or video that is generated in whole or substantially with the use of artificial intelligence." Unpaid content is outside that definition.

The format rules depend on the medium:

• Graphic ads: letters at least as large as the majority of the text, in the ad's language.
• Audio ads: spoken clearly at the beginning or end, for at least 3 seconds, in the ad's language.
• Video ads with audio: shown for at least 4 seconds and spoken at the beginning or end for at least 3 seconds, in the ad's language.

A prerecorded phone message that expressly advocates for or against a clearly identified candidate or a ballot question, and is generated in whole or substantially by AI, has its own rule. Under MCL 169.247(7), as amended by Public Act 263, it must say: "This message was generated in whole or substantially by artificial intelligence."

A person who knowingly violates MCL 169.247 is guilty of a misdemeanor punishable by a fine of up to $1,000, imprisonment for up to 93 days, or both.

A violation of MCL 169.259, the ad rule, is a state civil infraction, not a crime. The fine is up to $250 for a first violation and up to $1,000 for each second or later violation, and each noncompliant ad distributed or aired counts separately.

The section has exceptions for bona fide news coverage that acknowledges the content, stations paid to air the ad, certain news platforms with a clearly stated written policy, satire or parody, and businesses regulated by the Michigan Public Service Commission or the Federal Communications Commission.

Michigan Deepfake Laws: Election Media and Intimate Images (HB 5144, HB 4047)

MCL 168.932f, added by HB 5144 (Public Act 265 of 2023), bars distributing, or agreeing with another person to distribute, AI-produced materially deceptive media within 90 days before an election. It applies only when the person knows the media falsely represents the depicted individual and intends both to harm a candidate and to deceive voters, and a compliant disclaimer takes the media outside the ban.

A first violation is a misdemeanor punishable by up to 90 days, a fine of up to $500, or both. A violation within 5 years of a previous conviction is a felony punishable by up to 5 years, a fine of up to $1,000, or both.

The Protection from Intimate Deep Fakes Act (HB 4047, Public Act 11 of 2025) gives the person depicted a civil action and sets a misdemeanor of up to 1 year, a fine of up to $3,000, or both, rising to a felony of up to 3 years, a fine of up to $5,000, or both on listed aggravating facts. HB 4048 (Public Act 12 of 2025) only amends the sentencing guidelines.

Elements, defenses and remedies are in our Michigan deepfake laws guide.

AI Note-Takers and Michigan Recording Law

Whether an AI note-taker may record a Michigan meeting turns on the state's eavesdropping statute, MCL 750.539c. Our Michigan AI meeting recording guide covers it.

SB 760: The Michigan Chatbot Bill for Minors (LEAD for Kids Act)

SB 760, sponsored by Senators Polehanki and Geiss, passed the Senate 20-17 on April 29, 2026. The House referred it to its Communications and Technology committee the same day, and that was the last recorded action as of November 4, 2026.

The Senate-passed text would be the "leading ethical AI development for kids act." It covers an "advanced chatbot," meaning "a generative artificial intelligence system with natural language interface that provides ongoing, adaptive responses to user inputs, including via writing or sound." It excludes customer service bots, research or technical assistance systems, internal business tools, systems used in a clinical setting under a health professional's direct supervision, and narrowly tailored educational tools.

Under section 5, an operator could not make these available "unless the chatbot operator has actual knowledge that the user is not a minor":

• Features that simulate companionship or an interpersonal relationship
• Outputs that endorse, promote or facilitate suicide, self-harm, substantial physical harm to others, disordered eating, or unlawful drug or alcohol use
• Outputs that encourage secrecy, self-isolation, or not seeking help from a health professional or appropriate adult
• Outputs that optimize engagement in a way that supersedes the chatbot's safety or guardrails
• Outputs that are, describe or facilitate sexually explicit conduct or child sexual abuse material

A user's statement of being an adult would not be enough alone. The bill would also bar training a model on a minor's inputs without written consent from a parent or guardian.

Enforcement in section 7 runs two ways. The Attorney General could seek "a civil fine of $25,000.00 per violation," with each day a separate violation, plus injunctive relief and other remedies. A minor who suffers actual harm, or a parent or guardian on the minor's behalf, could sue for actual and punitive damages, among other relief.

Sections 4 and 5 would not apply until the effective date of the rules the Attorney General must first issue under section 19. So even if the bill were enacted, those duties would not begin on enactment.

HB 4668: The AI Safety and Security Transparency Act

HB 4668, introduced by Rep. Sarah Lightner on June 24, 2025, has moved between House committees and was re-referred to Communications and Technology on March 19, 2026. It has not passed the House. The details below come from the text as introduced.

A "large developer" must meet both tests: it developed a foundation model using computing power costing at least $5,000,000, and, within the preceding 12 months, one or more foundation models using computing power costing at least $100,000,000 in total.

"Critical risk" means a foreseeable and material risk of "the death of, or serious injury to, more than 100 people," or "more than $1,000,000,000.00 in damages to rights in money or property," through the kinds of incident the bill lists. Those are chemical, biological, radiological or nuclear weapons, cyberattacks, and certain conduct by a model with limited human intervention, along with harm by an intervening individual when the developer's activities made it substantially easier or more likely.

Large developers would have to:

• Produce, follow and conspicuously publish a safety and security protocol
• Publish a transparency report at least once every 90 days
• Retain a reputable third-party auditor at least once a year
• Refrain from retaliating against employees who report critical risks to authorities

Section 13 lets the Attorney General seek a civil fine of up to $1,000,000 per violation of section 7 or 9, plus injunctive or declaratory relief. A separate civil fine of up to $500 applies to a violation of section 11, the employee-protection section. An employee alleging retaliation could also bring a civil action within 90 days, and the Attorney General could seek an injunction if a developer's activities present an imminent critical risk. The introduced text starts its duties "on January 1, 2026," a date that has passed with the bill unenacted.

HB 5579 and SB 1077: Employer AI and Workplace Monitoring Bills

HB 5579 (Rep. Tsernoglou) was introduced February 24, 2026 and sent to the Economic Competitiveness committee, with no later action. SB 1077 (Sen. Camilleri) carries the same text and was introduced June 24, 2026 and sent to the Senate Labor committee. Neither is law.

The introduced text would be the "responsible artificial intelligence security for employees act." It would:

• Bar employers from using an automated decisions tool for an employment-related decision, except to screen large volumes of job applications for set hiring criteria or job skills
• Limit electronic monitoring to listed purposes and require written notice and written consent
• Limit retention of a covered individual's data to 3 years after the tool's purpose is achieved, unless a collective bargaining agreement says otherwise
• Bar selling or licensing that data
• Require an impact assessment before use
• Require notice of a security breach within 48 hours of discovery

Section 15 would allow a civil action by an aggrieved employee or applicant, or by someone acting for them, including a labor organization, plus a civil fine of up to $500 that a county prosecutor or the Attorney General could collect.

HB 4536 and HB 4537: AI in Claims Decisions

HB 4536 (Rep. Rheingans) would add a section to the Insurance Code. Its text says an insurer issuing or renewing a health insurance policy in Michigan "shall not deny, modify, or delay a claim based on a review using artificial intelligence."

HB 4537 would add a matching rule to the Social Welfare Act. Its text says the department or a contracted health plan "shall not deny, modify, or delay a claim under the medical assistance program based on a review using artificial intelligence." Both were re-referred to Communications and Technology on March 19, 2026. Neither is law.

DIFS Bulletin on Insurers' Use of AI

Separately from those bills, the Michigan Department of Insurance and Financial Services (DIFS) said in a release dated August 7, 2024 that it "has issued a new bulletin" on insurers' use of AI when they "underwrite new policies, make coverage determinations, or decide how much their products will cost."

The release says the bulletin "clarifies DIFS' expectations and outlines the responsibilities that insurers have" and "details how existing state laws apply to these technologies." As the release describes it, the bulletin is a regulator's statement of expectations under existing law, not a new law.

This page relies on the release for that description, so it does not say how the bulletin would apply to a particular claim denial.

HB 5899: Michigan's Proposed State AI Pilot Program

HB 5899 (Rep. Greene), introduced April 23, 2026, was reported on May 19, 2026 with a recommendation for referral to the Rules committee. The introduced text would create a three-member artificial intelligence governing board and a generative AI pilot program for state employees, agencies and departments, run by the Department of Technology, Management, and Budget.

The bill says the program could not violate the law, allow unauthorized access to personal data, adversely affect privacy rights, or discriminate in violation of the law. It is about state government, not private companies.

Other Michigan AI Bills

General Michigan Laws Without AI Language

Two general statutes are worth knowing, though neither names AI. The Elliott-Larsen Civil Rights Act, MCL 37.2202, bars employment discrimination because of religion, race, color, national origin, age, sex, sexual orientation, gender identity or expression, height, weight, or marital status. The Michigan Consumer Protection Act, MCL 445.903, makes listed unfair, unconscionable or deceptive practices in trade or commerce unlawful.

This page does not predict how a court would apply either one to an AI system. We did not review Attorney General enforcement files or court dockets for this page.

Federal AI Policy and Michigan

As of November 4, 2026, we found no federal statute in the public laws of the 119th Congress that sets general rules for private AI systems or generally preempts state AI laws. In December 2025, Executive Order 14365 directed federal agencies to challenge or discourage state AI laws the administration views as conflicting with national policy, and called for a federal framework from Congress. The order itself does not declare any state law invalid.

The Justice Department formed an AI Litigation Task Force in January 2026, the White House issued legislative recommendations in March 2026, and the FTC proposed a policy statement in July 2026. Separately, in April 2026 the Justice Department intervened in a lawsuit against Colorado's AI law under the Civil Rights Act of 1964. We did not find a published Commerce evaluation of state laws in the Federal Register or on ntia.gov, or a BEAD notice carrying the order's AI condition on NTIA's sites.

No case on our preemption tracker involves a Michigan law. See our pages on Trump's AI executive orders and state AI law preemption.

What Changes in Michigan AI Law in 2027

As of November 4, 2026, we found no AI law scheduled to take effect in Michigan in 2027. We searched the full text of the Michigan Compiled Laws and the Legislature's public act search for AI terms, and a law that uses other words would not surface in that search.

What could still change it is the current session. The Legislature is still meeting, with both chambers adjourned to November 5, 2026, and these matters are open:

• SB 760 (chatbots and minors): Awaiting House committee action since April 29, 2026. If enacted, sections 4 and 5 would wait for Attorney General rules.
• HB 4668 (frontier AI safety): In House committee since March 19, 2026.
• HB 5899 (state AI pilot): Its text says the pilot "must begin not later than January 1, 2027."
• HB 5579 and SB 1077 (workplace AI): In committee in each chamber.
• HB 4536 and HB 4537 (AI in claims decisions): In House committee since March 19, 2026.
• HB 6392 and HB 6417: Introduced September 24, 2026 and in committee.

Article IV, section 13 of the Michigan Constitution carries pending bills over only from a regular session held in an odd-numbered year. So 2025-26 bills not enacted by the final adjournment of the 2026 session would not carry into 2027, and a sponsor would have to introduce them again.

More Michigan Laws

Looking for information on other Michigan laws? Visit our AI Laws by State hub to compare Michigan with other states. You can also explore related topics:

• Michigan Surveillance Camera Laws for monitoring and recording rules
• Michigan Background Check Laws for employment screening regulations
• Michigan Recording Laws for wiretap and consent rules
• Michigan Deepfake Laws for election and intimate-image rules
• Michigan AI Meeting Recording Laws for AI note-takers

Frequently Asked Questions

Does Michigan have an AI law?

Yes, targeted ones. As of November 2026, the main ones we found are the 2023 election package (Public Acts 263-266), which covers AI disclosure in paid political ads and deceptive election media, and the Protection from Intimate Deep Fakes Act (Public Act 11 of 2025). We found no comprehensive AI act.

Does Michigan regulate AI chatbots?

Not by a chatbot statute we found. SB 760 would restrict companion features and harmful outputs for minors, but as of November 4, 2026 it has passed only the Senate, 20-17, and is in a House committee. This answer covers statutes and bills only.

What is SB 760 in Michigan?

SB 760 is the Leading Ethical AI Development for Kids Act. The Senate-passed text would bar chatbot operators from offering companion features and certain harmful outputs unless they have actual knowledge the user is not a minor, with Attorney General fines of $25,000 per violation and a private right of action for a harmed minor. It is not law.

What is HB 4668 in Michigan?

HB 4668 is the Artificial Intelligence Safety and Security Transparency Act, a bill requiring large foundation-model developers to publish safety protocols and transparency reports and obtain annual third-party audits. As of November 4, 2026, it is in a House committee and has not passed the House.

Does Michigan regulate AI in hiring?

We found no Michigan statute written specifically for AI in hiring. HB 5579 and SB 1077 would restrict employer use of automated decision tools and electronic monitoring, but neither has been enacted. MCL 37.2202 bars employment discrimination in general and does not name AI. This answer covers statutes and bills only.

Can Michigan employers use AI to monitor employees?

We found no enacted Michigan statute aimed at AI monitoring of workers. HB 5579 and SB 1077 would limit monitoring to listed purposes and require written notice and consent, but they are bills in committee as of November 4, 2026. This answer covers statutes and bills only.

Do Michigan political ads have to disclose AI?

Paid ones with AI-generated image, audio or video do. Since February 13, 2024, MCL 169.259 has required a statement that the ad was generated in whole or substantially by AI. A violation is a civil infraction with a fine of up to $250 for a first violation and up to $1,000 for each later violation.

Are AI deepfakes illegal in Michigan?

Some are. MCL 168.932f covers distributing, or agreeing to distribute, AI-produced deceptive election media within 90 days before an election, on knowledge and intent conditions. The Protection from Intimate Deep Fakes Act covers realistic sexual deep fakes of identifiable people. Michigan has no general ban on all deepfakes that we found.

What is the penalty for an intimate deepfake in Michigan?

Under section 8 of the Protection from Intimate Deep Fakes Act, a violation is a misdemeanor of up to 1 year, a fine of up to $3,000, or both. On listed facts, such as posting the deep fake on a website or a prior conviction, it is a felony of up to 3 years, a fine of up to $5,000, or both. The person depicted may also sue.

Can Michigan insurers use AI to deny claims?

We found no enacted Michigan statute written specifically for this. HB 4536 would bar an insurer from denying, modifying or delaying a health insurance claim based on a review using AI, but as of November 4, 2026 it is in a House committee. Separately, Michigan's insurance regulator, DIFS, announced a bulletin on August 7, 2024 stating its expectations for insurers' AI use, including in coverage determinations, and how existing state laws apply. The bulletin is not a new law.

Does Michigan have any AI law taking effect in 2027?

We found none as of November 4, 2026. The session is still running, so a bill enacted before it ends could change that. Bills not enacted in 2026 do not carry over and would have to be introduced again.