Michigan Deepfake Laws: AI Images, Voice Cloning & Penalties (2026)

Michigan has enacted laws targeting two of the three deepfake buckets: nonconsensual intimate deepfakes and election deepfakes. Michigan's intimate-image deepfake statute took effect August 26, 2025, and its election-media rules have been in force since February 13, 2024. Voice cloning and digital likeness protection remain limited to common law only. This page covers Michigan-specific deepfake statutes; for the full national picture, see Deepfake & AI Voice Cloning Laws by State.

Is It Illegal to Make a Deepfake of Someone in Michigan?

It depends on the category of deepfake. Michigan law now covers two of the three main buckets: sexual and intimate deepfakes (adults), and election disinformation deepfakes. The third bucket, AI voice cloning and digital likeness outside those contexts, has no dedicated state statute.

Creating a deepfake for satire, entertainment, or commentary on a matter of public concern is not automatically illegal in Michigan, provided it does not involve intimate imagery or election manipulation. That said, the election law carries a significant First Amendment caveat: enforcement of election deepfake statutes nationwide has faced court challenges, and the legal landscape is still developing. The Michigan law includes a safe harbor for depictions that carry a proper disclaimer, which mitigates constitutional risk.

Deepfakes used for fraud, extortion, or harassment may also be prosecuted under Michigan's existing fraud, extortion, or cyberstalking statutes independent of the deepfake-specific laws. If a fake video or cloned voice was used to trick you into paying someone, see what to do after a scam for the first steps to take and where to report it.

Sexual and Intimate Deepfakes

Michigan enacted comprehensive intimate deepfake law in August 2025 when Governor Whitmer signed PA 11 of 2025 (the Protection from Intimate Deep Fakes Act). The criminal statute targets individuals who intentionally create or disseminate a deepfake depicting intimate body parts or sexual acts of an identifiable person when the creator knew or should have known the depiction would cause the person harm.

The base offense is a misdemeanor carrying up to 1 year in jail and a $3,000 fine. The offense escalates to a felony (up to 3 years, $5,000) when any aggravating factor is present: the victim suffers financial loss, the creator intended to profit, the creator maintained a website or app for creating or sharing deepfakes, the deepfake was posted online, the creator intended to harass, extort, or threaten the victim, or the creator has a prior conviction under the act. PA 12 of 2025 incorporated the felony into Michigan's sentencing guidelines as a Class F offense.

Consent is not a defense unless the depicted person signed a written, plain-language agreement that includes a general description of the intimate depiction. That is a high bar intentionally designed to prevent coerced or vague consent from defeating prosecution.

PA 11 also creates a civil cause of action. Victims can seek economic and noneconomic damages (including compensation for mental anguish, embarrassment, and humiliation), disgorgement of the defendant's profits, attorney fees and court costs, and temporary or permanent injunctive relief. Civil fines of up to $1,000 per day accrue for injunction violations.

For minors, AI-generated child sexual abuse material is covered separately under MCL 750.145c, which expressly includes "computer-generated image" in its definition of child sexually abusive material. There is no gap in Michigan law for AI-generated CSAM: the statute covers any depiction that "appears to include a child" engaging in listed sexual acts, regardless of whether a real child was used in production.

Election and Political Deepfakes

MCL 168.932f, added by 2023 PA 265 and effective February 13, 2024, prohibits distributing materially deceptive media within 90 days before any election when the distributor (a) knows the media falsely represents a depicted individual, (b) intends the distribution to harm a candidate's reputation or electoral prospects, and (c) intends the distribution to deceive voters about whether the depicted individual actually engaged in the depicted speech or conduct. All of these elements must be present together, not just one.

The statute defines "materially deceptive media" as an AI-produced image, audio recording, or video that falsely depicts someone engaging in speech or conduct they did not actually perform and that would cause a reasonable viewer or listener to believe the depiction is real. That definition squarely covers synthetic audio, video, and still images generated by AI.

The disclaimer safe harbor is meaningful: if the deepfake includes a clear statement that the media was manipulated by technical means, with formatting requirements tailored to the medium (text overlay for video, spoken disclosure for audio, visible text for images), distribution is lawful even within the 90-day window. This is intended to preserve space for political satire and commentary.

Penalties escalate sharply for repeat offenders. A first violation is a misdemeanor (up to 90 days, $500 fine). A second violation within five years is a felony carrying up to 5 years and a $1,000 fine. The Michigan Attorney General, the depicted individual, injured candidates, and voter advocacy organizations may all seek permanent injunctive relief.

Michigan passed a second AI election law the same day, and it operates on a completely different trigger. MCL 169.259, added to the Michigan Campaign Finance Act by 2023 PA 263 and effective February 13, 2024, requires that any "qualified political advertisement" carry a clear and conspicuous statement that it was generated in whole or substantially by artificial intelligence. A qualified political advertisement is any paid advertisement relating to a Michigan candidate, election, or ballot question that contains image, audio, or video "generated in whole or substantially with the use of artificial intelligence." Unlike MCL 168.932f, this is a flat disclosure mandate on the advertiser: there is no intent-to-deceive element and no 90-day window, so it applies to an AI-generated political ad run at any point in the cycle.

The disclosure format rules are specific to the medium. In a graphic or text advertisement, the statement must appear in letters at least as large as the majority of the other text and in the same language as the ad. In an audio advertisement, it must be spoken clearly and intelligibly at the beginning or the end for at least 3 seconds. In a video advertisement with audio, it must both appear on screen for at least 4 seconds in letters at least as large as the majority of any other text and be spoken for at least 3 seconds. A first violation is a civil infraction carrying a fine of up to $250, and each subsequent violation carries a fine of up to $1,000, assessed separately for each advertisement distributed or aired. The statute carries exemptions, including for news coverage and for satire or parody.

A broader constitutional note: in August 2025 a federal judge in the Eastern District of California held California's election deepfake statute (AB 2839) facially unconstitutional under the First Amendment, found no portion of it severable, and permanently enjoined the state from enforcing it against the named plaintiffs in that case (Kohls v. Bonta, 797 F. Supp. 3d 1177 (E.D. Cal. Aug. 29, 2025)). The injunction runs to those plaintiffs rather than operating as a statewide bar on all enforcement. Michigan's laws have not faced a similar ruling as of mid-2026, but election deepfake laws nationwide carry ongoing First Amendment litigation risk. The disclaimer safe harbor in MCL 168.932f is Michigan's primary structural protection against such a challenge.

AI Voice Cloning and Digital Likeness

Michigan has no statutory right of publicity and no ELVIS Act-style law protecting voice or likeness against AI replication. Michigan courts recognize an appropriation tort under common law, treating it as a property-right claim, but that doctrine predates AI voice cloning and has not been extended by statute to cover synthetic voice replicas.

By contrast, Tennessee's ELVIS Act (Tenn. Code Ann. 47-25-1101 et seq., effective July 1, 2024) is the national archetype: it expressly extends the right of publicity to voice and requires affirmative consent before an AI can simulate a person's voice for commercial purposes. Michigan offers none of that protection at present.

Practically, a Michigan performer or public figure whose voice is cloned by AI for commercial use without permission has limited state-law remedies. Common law appropriation might support a claim if commercial exploitation is clear, but the outcome is uncertain. Federal law fills some of the gap: the FTC's Impersonation Rule (16 CFR Part 461, effective April 1, 2024) prohibits AI voice impersonation of government entities and businesses, and the FCC has ruled that AI-generated voices in robocalls are illegal under the TCPA. But these rules do not protect private individuals from having their voices cloned in non-robocall contexts.

For general AI regulation in Michigan beyond deepfake-specific rules, see the Michigan AI Laws page, which covers the state's broader approach to AI governance, automated decision-making, and emerging AI legislation.

Federal Law That Applies in Michigan

Several federal laws apply to Michigan residents and platforms regardless of what state law says.

The TAKE IT DOWN Act (Public Law 119-12, signed May 19, 2025) is the first federal intimate deepfake law. It makes it a federal crime to knowingly publish nonconsensual intimate visual depictions of adults or minors, expressly including AI-generated "digital forgeries." Penalties reach up to 2 years in prison (3 years when minors are involved). Platforms must remove flagged content within 48 hours of a victim's notice request, with the compliance deadline having taken effect May 19, 2026. The FTC enforces the platform-removal obligation.

The DEFIANCE Act (S.1837, 119th Congress) is a separate federal bill that would create a civil cause of action for sexual deepfake victims with liquidated damages of $150,000 (or $250,000 if the conduct involved actual or attempted sexual assault, stalking, or harassment). The Senate passed S.1837 by unanimous consent on January 13, 2026, but as of mid-2026 the bill remains pending in the House and is not law. It is not yet enacted law. For background on the bill, see DEFIANCE Act: Deepfake Porn Victims Can Sue for $150K+.

The NO FAKES Act (S.4591, 119th Congress), which would create a federal right of publicity against unauthorized AI voice and likeness replicas, is also proposed only and has not passed either chamber as of June 2026.

Federal CSAM law (18 U.S.C. 2256(8)(B)) covers AI-generated images "indistinguishable" from a real minor, closing any gap that might exist in state coverage. The FCC's February 2024 ruling (FCC 24-17) makes AI-generated voices in robocalls illegal under the TCPA nationwide.

What Victims Can Do

A Michigan victim of an intimate deepfake has several parallel routes for relief.

On the criminal side, victims should report to local law enforcement or the Michigan Attorney General's office. PA 11 of 2025 is the primary criminal statute; if the deepfake was created or shared online, federal authorities (FBI) can investigate TAKE IT DOWN Act violations simultaneously.

On the civil side, PA 11 gives Michigan victims a direct cause of action for economic damages, emotional distress damages, disgorgement of the defendant's profits, and injunctive relief. A victim does not need to wait for a criminal prosecution to file a civil lawsuit.

For platform takedowns, the TAKE IT DOWN Act requires covered platforms to remove flagged intimate deepfakes within 48 hours of a victim's notice. This is separate from and faster than most platforms' voluntary content-moderation processes. Victims should submit takedown requests directly through each platform's notice mechanism and simultaneously report to the FTC if platforms fail to comply within the 48-hour window.

For election deepfakes, the depicted individual, the AG, injured candidates, and voter organizations can seek injunctive relief under MCL 168.932f. Given the 90-day pre-election window, timing matters: emergency TRO applications may be necessary for rapid relief. A missing AI disclosure on a paid political ad is a separate matter handled as a civil infraction under MCL 169.259.

Penalties at a Glance

Disclaimer: This page provides general legal information about Michigan deepfake laws and is not legal advice. Deepfake laws are a fast-moving area of law; statutes, court interpretations, and pending federal legislation can change quickly. If you have been harmed by a deepfake or face a legal matter in this area, consult a licensed Michigan attorney.

Sources

Citations for this page are listed below.

For more on Michigan recording law generally, see the Michigan Recording Laws page. For data protection rights in Michigan, see Michigan Data Privacy Laws.

More Michigan Laws

• Michigan AI Meeting Recording Laws
• Michigan Data Privacy Laws
• Michigan Recording Laws
• Michigan Sexting Laws

Frequently Asked Questions

Is it illegal to make a deepfake of someone in Michigan?

It depends on the type of deepfake. Creating or sharing an intimate deepfake of an identifiable person is a criminal offense under Michigan PA 11 of 2025, punishable by up to 1 year (misdemeanor) or 3 years (felony with aggravating factors). Creating AI-generated election disinformation within 90 days of an election is prohibited under MCL 168.932f, and a paid political ad generated by AI must carry an AI disclosure under MCL 169.259 at any time of year. Deepfakes for satire or commentary that do not involve intimate imagery or election manipulation are generally not prohibited, though other laws such as defamation or fraud may still apply.

Is deepfake porn a crime in Michigan?

Yes. Michigan PA 11 of 2025 (the Protection from Intimate Deep Fakes Act) makes it a crime to intentionally create or share a deepfake depicting intimate body parts or sexual acts of an identifiable person without consent, when the creator knew or should have known it would cause harm. The base offense is a misdemeanor (up to 1 year, $3,000). Aggravating factors, such as posting online or intent to harass or extort, elevate the charge to a felony (up to 3 years, $5,000).

Can I sue someone for making a deepfake of me in Michigan?

Yes, for intimate deepfakes. PA 11 of 2025 creates a civil cause of action allowing victims to recover economic damages, noneconomic damages (including mental anguish and humiliation), the defendant's profits, attorney fees, and injunctive relief. Civil fines of up to $1,000 per day apply for injunction violations. For other types of deepfakes not covered by PA 11, existing tort claims such as defamation, false light, or intentional infliction of emotional distress may be available depending on the facts.

Are AI voice clones illegal in Michigan?

Michigan has no statute that directly prohibits AI voice cloning of private individuals. There is no ELVIS Act-style right-of-publicity law in Michigan. Common law recognizes an appropriation tort, but it has not been extended by courts or legislature to cover synthetic AI voice replicas. If the voice clone is used in robocalls, it is illegal nationwide under the FCC's February 2024 TCPA ruling. If used to impersonate a government entity or business, the FTC's Impersonation Rule applies.

Does Michigan law cover AI-generated child sexual abuse material?

Yes. MCL 750.145c expressly includes 'computer-generated image' in its definition of child sexually abusive material. The statute covers any depiction that appears to include a child engaging in listed sexual acts, regardless of whether a real child was involved in production. Federal law (18 U.S.C. 2256) provides an additional layer of coverage for AI-generated CSAM indistinguishable from a real minor.

What is the 90-day election deepfake rule in Michigan?

MCL 168.932f prohibits distributing AI-generated media that falsely depicts a candidate or political figure within 90 days before an election, when done with the intent both to harm a candidate's reputation or electoral prospects and to deceive voters about whether the depicted individual actually said or did what is shown. A disclaimer safe harbor applies: if the media clearly discloses that it was manipulated by technical means, with format-specific requirements for video, audio, and images, distribution is lawful even within the 90-day window.

Do AI-generated political ads have to be labeled in Michigan outside the 90-day window?

Yes. MCL 169.259, added by 2023 PA 263, requires any qualified political advertisement generated in whole or substantially by artificial intelligence to carry a clear and conspicuous statement saying so. It applies year-round, with no 90-day window and no intent-to-deceive element, which makes it broader than MCL 168.932f. The disclosure must run at least 3 seconds spoken in audio, at least 4 seconds on screen in video (plus the 3-second spoken statement), and in graphics must use letters at least as large as the majority of the other text. A first violation is a civil infraction with a fine of up to $250; each later violation carries up to $1,000.

How does the federal TAKE IT DOWN Act work in Michigan?

The TAKE IT DOWN Act (Public Law 119-12, signed May 19, 2025) applies in all 50 states including Michigan. It makes it a federal crime to publish nonconsensual intimate visual depictions, including AI deepfakes, and requires platforms to remove flagged content within 48 hours of a victim's notice. Victims can report violations to the FTC. The federal law works alongside Michigan's PA 11 of 2025, giving victims both state and federal remedies simultaneously.

Is the DEFIANCE Act law in Michigan?

No. The DEFIANCE Act (S.1837) is a proposed federal bill that would create a civil cause of action for sexual deepfake victims with liquidated damages of $150,000, or $250,000 if the conduct involved actual or attempted sexual assault, stalking, or harassment. The Senate passed S.1837 by unanimous consent on January 13, 2026, but as of mid-2026 the bill remains pending in the House and has not been enacted. Michigan victims of intimate deepfakes already have civil remedies under state law through PA 11 of 2025.