Election Deepfake Laws by State: Labels, Bans, Lawsuits (2027)

Election deepfakes are illegal only in some states and only within each law's limits: as of November 2026, we found no federal statute written for them, and by our reading of each state's statutes as of November 5, 2026, 31 states have a law written for them, of which 24 require a label or disclosure, 4 bar covered fakes for most speakers within set limits, and 3 give a depicted candidate a lawsuit rather than a ban or label duty.

Status as of November 5, 2026: No federal appeals court ruling on whether an election deepfake law violates the First Amendment, in the cases we found. California rewrote its main election deepfake law in September 2026, and we found no ruling on the new text. Laws in Hawaii, Montana and Vermont are under court orders of different scope, and no election deepfake bill has passed Congress.

Are Election Deepfakes Illegal?

Some are, in some states, and only within each law's limits. Most state election deepfake laws are disclosure rules, so a fake that carries the required label may be lawful under that statute. A handful bar covered fakes outright for most speakers, and three states only give the depicted candidate a way to sue.

Every one of these laws has conditions. Common ones are a pre-election window, knowledge that the media is false, an intent to injure a candidate or mislead voters, and media realistic enough that a reasonable person would believe it. A fake that misses one condition can fall outside the statute even in a state with a law.

Not every election deepfake law is AI-only, and not every one reaches still images. Texas covers video only, and Kentucky and Idaho reach audio or video. Read the state page before assuming a law fits a given image, clip or call.

How Many States Have Election Deepfake Laws?

By our reading of each state's statutes as of November 5, 2026, 31 states have a law written for election deepfakes. We sort them by what the law does when a covered fake appears:

The National Conference of State Legislatures, on a page it says was updated October 13, 2026, states: "Thirty-one states have enacted laws regulating the use of deepfakes in political messaging." NCSL uses its own criteria and we did not compare its list with ours, so the matching number should not be read as the same 31 states.

Two of our categories need a note. Kentucky is counted as a civil-suit state because its first remedy is the candidate's suit, but its statute also routes disobedience of the court's disclosure order to a felony penalty. Mississippi's and Montana's state pages describe their laws as bans with a label defense or safe harbor. We count them with the label laws because a compliant disclosure takes the communication outside Montana's ban and is a defense to prosecution in Mississippi; whether a labeled fake can still be enjoined in Mississippi is not settled by the text.

Election Deepfake Laws in All 50 States and DC

Each state links to its full page, which is the source for its row. "No ruling found" means the state page searched and found none; "None tracked" means the law did not come up in our litigation review, not that no case exists. Windows and penalties are each statute's own terms.

Label Laws: What a Disclosure Does and Does Not Do

In most label states, the disclosure is the line between lawful and unlawful under that statute. In Mississippi it is a defense to prosecution, and the statute's injunction route is not tied to it. Each state writes its own wording, and most set format rules: text no smaller than other text, shown for a whole video, or spoken at the start and end of audio.

The wording differs from state to state. Delaware's is "This (image/video/audio) has been altered or artificially generated." Florida's is "Created in whole or in part with the use of generative artificial intelligence (AI)." North Dakota's is "THIS CONTENT GENERATED BY ARTIFICIAL INTELLIGENCE."

Who carries the duty differs too. Indiana, Utah and Wisconsin put it on paid communications or the payer, Montana on persons working in an official election capacity, and Rhode Island on candidates, committees and independent spenders. New York's applies year-round to anyone who distributes or publishes with actual knowledge.

A label does not settle everything. Complying with Wisconsin's rule does not excuse other liability, and a sexual deepfake of a candidate can still be a crime under a state's intimate-image law, as Idaho's page notes for Idaho Code 18-6606.

Bans: California, Maryland, Minnesota and Texas

California. AB 502 (Stats. 2026, ch. 333), approved September 19, 2026 as an urgency statute, rewrote Elections Code 20012. It bars knowing distribution, with malice, of materially deceptive content in the 30 days before an election for candidates and elected officials, and from 30 days before through 38 days after for elections officials and election items. Its disclosure routes are narrow: a candidate portraying themselves, bona fide broadcast news, and satire or parody.

Maryland. Election Law 16-905, in force June 1, 2026, bars knowingly or with reckless disregard creating, using or disseminating a deepfake to produce materially false information, with a listed voting-related intent and actual or intended harm. It has no window, and the enacted text gives no defense for adding a label.

Minnesota. Section 609.771 reaches a deep fake disseminated without the depicted person's consent, with intent to injure a candidate or influence an election, by one who knows or acts with reckless disregard about whether it is a deep fake. The windows are 90 days before a party nominating convention, or after absentee voting starts.

Texas. Election Code 255.004(d) needs both creating a deep fake video and causing it to be published or distributed within 30 days of an election, with intent to injure a candidate or influence the result. Audio and still images are outside its definition.

Civil-Suit States: Idaho, Kentucky and Washington

Idaho. Idaho Code 67-6628A lets a candidate deceptively represented by realistic but false audio or video in an electioneering communication sue for an injunction and damages. A "This (video/audio) has been manipulated" disclosure is an affirmative defense, not a requirement, and we found no fine or jail term tied to the section.

Kentucky. KRS 117.322 lets a candidate whose appearance, action or speech is altered by synthetic media, without consent, in an electioneering communication within 45 days of a primary or regular election, sue the sponsor for an order requiring a clear and conspicuous disclosure. Failing to obey that order carries the penalties of KRS 121.990(3), where a knowing violation is a Class D felony.

That cross-reference is why Kentucky is a civil-suit state only in its first step. The section's own words route a violated order to the felony provision, and we found no ruling on how a court would apply it.

Washington. RCW 29B.35.010 gives a candidate a suit over synthetic media in an electioneering communication: state, local and judicial races, within 60 days of an election, at a value of $1,000 or more. The chapter says nothing in it is a violation under other chapters of the title, and we found no fine or crime in the sections we read.

Court Challenges to Election Deepfake Laws

Courts have limited several election deepfake laws, but the rulings are narrower than headlines suggest. Federal district courts have ruled against laws in California, Hawaii and Montana, with California's and Montana's relief limited to the parties who sued; a separate ruling held California's platform law preempted by Section 230, and that decision is on appeal. In Minnesota, the Eighth Circuit decided only that the challengers lacked standing or waited too long. As of November 5, 2026, we found no federal appeals court ruling on whether an election deepfake law violates the First Amendment.

Each entry below rests on the court document named, and its date is the newest document we read. We did not review live dockets, so later filings may exist.

• Kohls v. Ellison (Minnesota). On February 9, 2026, the Eighth Circuit, No. 25-1300, affirmed the denial of a preliminary injunction. It held that "Kohls's videos, labeled as parodies, are not deep fakes under the statute," and that a state legislator who reshared one had not explained her "sixteen-month delay." It noted that the district court "did not rule on the constitutionality of the statute."
• X Corp. v. Ellison (Minnesota). On December 2, 2025, the district court dismissed X Corp.'s Section 230 claim without prejudice for lack of standing and stayed the case. It did not reach the First Amendment claims.
• Kohls v. Bonta (California). A federal court permanently barred California's Attorney General and Secretary of State, and those working for or acting in concert with them, from enforcing the 2024 version of this law against three named plaintiffs: Christopher Kohls, The Babylon Bee and Kelly Chang Rickert. In September 2026 the Legislature rewrote it, narrowing the window to 30 days before an election, running through 38 days after it for elections officials and election items, and we found no court ruling on the new text. Whether that injunction reaches the rewritten section is an open question.
• AB 2655 (California platforms). The same court's August 20, 2025 judgment held the platform law preempted by Section 230 as applied to X Corp. and Rumble and to any such provider of an interactive computer service. It was not a First Amendment ruling. A stipulated order bars enforcement against any such provider. The appeal, Ninth Circuit No. 25-6138, had no decision we found as of November 5, 2026.
• The Babylon Bee, LLC v. Lopez (Hawaii). On January 30, 2026, the District of Hawaii, No. 1:25-cv-00234, held Act 191 content- and speaker-based, overbroad and vague, and ruled that "State Defendants are PERMANENTLY ENJOINED from enforcing Act 191." The order defines that term as six named officials, among them the Attorney General; whether the Honolulu prosecutor, named only in a footnote, is bound is not clear from the order. We found no appeal in two web searches.
• Accountability in State Government v. Knudsen (Montana). On September 16, 2026, the District of Montana, No. 6:26-cv-00038, granted a preliminary injunction, finding a likelihood of success on the First Amendment challenge. Citing Trump v. CASA, it limited relief to the two named plaintiffs, so the Act still applies to others.
• Poitras v. Clark (Vermont). On September 4, 2026, the District of Vermont, No. 2:26-cv-234, issued a narrow preliminary injunction barring the Attorney General's office from acting on one civil investigative demand or taking further Act 75 enforcement against the plaintiff over one video. It said it "does not consider that broad, facial challenge at this time."
• Ex parte Stafford (Texas). On September 4, 2024, the Texas Court of Criminal Appeals, PD-0310-23, held that "255.004(b) on its face, violates the First Amendment." Subsection (b) is the "true source" rule, not the deepfake subsection, so (d) was neither upheld nor struck down. Rehearing was denied October 30, 2025.

Federal Election Deepfake Law

No federal statute is written for election deepfakes, by our search, though existing federal laws can reach some of them. The Federal Election Commission voted on September 19, 2024 not to open a rulemaking on AI in campaign ads, and adopted an interpretive rule. It said fraudulent misrepresentation of candidates "may be accomplished using AI-assisted media" and that the statute and its regulation are "technology neutral" (52 U.S.C. 30124; 11 CFR 110.16).

That rule did not create a deepfake ban. As of November 5, 2026, the FEC's own site lists two commissioners, and it needs four votes to act.

The FCC proposed AI disclosure rules for broadcast political ads in 2024 (MB Docket No. 24-211, released July 25, 2024). We found no final rule.

In Congress, the Protect Elections from Deceptive AI Act (S. 1213, H.R. 5272), the AI Transparency in Elections Act of 2026 (H.R. 9922) and the REAL Political Advertisements Act (H.R. 10814) are among the election bills referred to committee, and none is law. Our federal deepfake bills page tracks each one.

Federal law separately makes it a crime to knowingly publish certain intimate images of identifiable people, including AI forgeries indistinguishable from real images, when the elements of the TAKE IT DOWN Act, 47 U.S.C. 223(h), are met, and requires covered platforms to remove reported depictions within 48 hours of a valid request (47 U.S.C. 223a). See our TAKE IT DOWN Act page.

AI robocalls are a separate track. The FCC's September 2024 order holding a political consultant liable for a $6,000,000 forfeiture over the January 2024 fake Biden robocalls rests on the Truth in Caller ID rule, which the order says is "not dependent on the content of the call." A New Hampshire jury acquitted him of all state charges in June 2025, the Associated Press reported. Our AI voice scam calls page covers call rules.

For Voters, Campaigns and Candidates

If you see a suspected election deepfake. Check the claim against the candidate's or election office's own channels before sharing it; under several of these laws, distributing a covered fake is the act the statute reaches. Some states name a place to report:

• Colorado: a written complaint to the Secretary of State (C.R.S. 1-46-104).
• Florida: a complaint to the Florida Elections Commission (Fla. Stat. 106.145).
• Maryland: the State Board's public reporting portal, which can refer matters to the State Prosecutor (Election Law 2-110(b)).
• Montana: the Commissioner of Political Practices, 406-444-2942 (MCA 13-35-804).
• New Hampshire: a written complaint to the attorney general (RSA 664:18).
• Oregon: a written complaint by an elector to a filing officer (ORS 260.345).
• Wisconsin: a verified petition to the Ethics Commission (Wis. Stat. 11.1400(6)).

If you run or pay for political ads. Read the rule for every state where the ad will run, since the wording, format and covered media differ. None of the statutes in Maryland (16-905), Minnesota (609.771) or Texas (255.004(d)) gives a defense for adding a label, though the Eighth Circuit read Minnesota's realism definition to exclude videos labeled as parodies, in a standing ruling on that record. California's disclosure routes are narrow.

Satire is not handled the same way everywhere. Alabama's exemption covers satire substantially dependent on physical or verbal impersonation, not on AI. New Mexico's ad rule protects satire only with the disclaimer, which a 2025 Attorney General opinion, an advisory view, called likely unconstitutional on its face. The Eighth Circuit said labeled parody videos were not deep fakes under Minnesota's statute, in a standing ruling on that record.

If you are the candidate depicted. Who may sue depends on the state: the depicted candidate in Idaho, Indiana, Kentucky, Rhode Island and Washington, and a wider set, including the Attorney General or voter groups, in Alabama, Michigan, Mississippi, New Mexico and South Dakota. Many of these suits require clear and convincing evidence. A state bar lawyer referral service can point you to an election lawyer, and the state pages list each statute's remedies.

What Changes in Election Deepfake Law in 2027

Dated events we found, each from the state or federal page that sources it:

• January 1, 2027, California. AB 686 (Stats. 2026, ch. 334) moves the repeal date of Elections Code 20010, the 60-day label law, from January 1, 2027 to January 1, 2031. It adds no new rule.
• January 1, 2027, Louisiana. Act 773 of 2026 (SB 495), a campaign finance law, takes effect. It does not amend the AI disclosure rule, but it may change how the subsections of R.S. 18:1463 are lettered.
• January 3, 2027, Congress. A bill that is not enacted before the 119th Congress ends on January 3, 2027 does not carry over. It would have to be introduced again in the next Congress. That includes every federal election deepfake bill named above.
• New York: no 2027 ban. A 90-day election deepfake ban appeared in early prints of the 2026 budget bill, but the enacted print, chapter 55 of 2026, lists that part as intentionally omitted. Election Law 14-106(5) remains a disclosure rule.

Matters with no date yet: the Kohls v. Ellison remand, the Ninth Circuit's AB 2655 appeal, any ruling on California's rewritten 20012, the next steps in the Montana and Vermont cases, the Colorado complaint hearing, and the Massachusetts H.5700 budget talks. Hawaii's Act 191 window would next open on the first working day of February 2028.

New legislative sessions begin in 2027, and we cannot say whether any election deepfake bill will pass. Illinois's and North Carolina's new General Assemblies convene January 13, 2027, according to those state pages.

Related Deepfake Laws

• Deepfake laws by state
• Federal deepfake bills
• AI voice cloning laws
• AI voice scam calls
• Deepfake fraud and impersonation laws
• Right of publicity
• TAKE IT DOWN Act
• AI laws

Frequently Asked Questions

Are election deepfakes illegal?

Sometimes, depending on the state. As of November 5, 2026, we found no federal statute written for election deepfakes. By our reading, 31 states have a law written for them: 24 require a label or disclosure, so a deceptive fake carrying the required disclosure may be allowed; 4 (California, Maryland, Minnesota and Texas) bar covered fakes for most speakers within set limits; and 3 (Idaho, Kentucky and Washington) give a depicted candidate a lawsuit. Each law has its own window and knowledge or intent elements, and we found no such law in 19 states and DC.

Is there a federal law against election deepfakes?

Not one written for them, as of November 5, 2026. The Federal Election Commission declined in 2024 to write AI rules and said its existing ban on fraudulent misrepresentation of candidates is technology neutral, so it can reach AI-assisted media used that way. Separately, the FCC in 2024 imposed a $6,000,000 forfeiture on the consultant behind the fake Biden robocalls, under its Truth in Caller ID rule, for spoofed caller ID used with intent to defraud and cause harm, not for the AI voice itself. The FCC's 2024 proposal on AI disclosure in broadcast political ads has no final rule we found. Bills such as S. 1213, H.R. 9922 and H.R. 10814 were referred to committee and are not law.

Which states ban election deepfakes?

By our reading as of November 5, 2026, four: California (Elections Code 20012, knowing distribution with malice within 30 days before an election, with narrow disclosure routes), Maryland (Election Law 16-905, knowingly or with reckless disregard creating, using or disseminating a deepfake to produce materially false information, with a listed voting-related intent and harm, no window), Minnesota (609.771, dissemination without consent, knowing or with reckless disregard that it is a deep fake, with intent to injure a candidate or influence an election, in set windows) and Texas (Election Code 255.004(d), creating a deep fake video and causing it to be published or distributed within 30 days, with intent to injure a candidate or influence the result).

Which states require a label on AI political ads or election deepfakes?

By our reading as of November 5, 2026, 24 states have label or disclosure laws for election deepfakes: Alabama, Arizona, Colorado, Delaware, Florida, Hawaii, Indiana, Louisiana, Maine, Michigan, Mississippi, Montana, Nevada, New Hampshire, New Mexico, New York, North Dakota, Oregon, Rhode Island, South Dakota, Tennessee, Utah, Vermont and Wisconsin. Each sets its own wording, covered media and timing, and Hawaii's, Montana's and Vermont's are limited by court orders of different scope.

Can a candidate sue over an election deepfake?

In many states, yes, if the statute's conditions are met. A depicted candidate can sue under the election deepfake laws of Idaho, Kentucky and Washington, which are built around that suit, and many label and ban states also let a depicted candidate seek an injunction, sometimes damages. Most of these suits require clear and convincing evidence, and in some states only an injunction is available.

Did a court strike down a state election deepfake law?

Not across the board. Federal district courts ruled against California's 2024 law, Hawaii's Act 191 and Montana's SB 25, but the California and Montana relief runs only to the parties who sued, and Hawaii's injunction binds the named State Defendants. A narrow preliminary injunction also bars Vermont's Attorney General's office from further enforcement against one plaintiff over one video. California rewrote its law in September 2026 and we found no ruling on the new text. As of November 5, 2026, we found no federal appeals court ruling on whether an election deepfake law violates the First Amendment.

Is AI satire or parody of a candidate allowed?

Most election deepfake laws exempt satire or parody, but the terms differ. Alabama's exemption covers satire that depends on physical or verbal impersonation, not AI. New Mexico's ad rule protects satire only with the AI disclaimer. California's rewritten law exempts content a reasonable person would understand as satire or parody, or that carries a stated satire disclosure.

Will New York ban election deepfakes in 2027?

We found no such ban. A 90-day ban appeared in early prints of New York's 2026 budget bill, but the enacted print, chapter 55 of 2026, lists that part as intentionally omitted. As of November 5, 2026, Election Law 14-106(5) remains a disclosure rule for political communications with materially deceptive media.

How many states have election deepfake laws?

By our reading of each state's statutes as of November 5, 2026, 31 states have a law written for election deepfakes: 24 label laws, 4 bans for most speakers and 3 civil-suit laws. The National Conference of State Legislatures, on a page updated October 13, 2026, also says 31 states have enacted laws on deepfakes in political messaging, but its list was not compared with ours, so the two may not name the same states.