Deepfake law went from nothing to everywhere in 18 months. The TAKE IT DOWN Act (signed May 19, 2025, Public Law 119-12) created the first federal criminal penalties for non-consensual intimate deepfakes. State legislatures have added deepfake statutes of their own. And a wave of civil litigation, from Taylor Swift's unauthorized AI images to AI-generated political ads, is building the case law framework in real time. For attorneys, deepfakes touch defamation, privacy, IP, election law, employment law, and criminal law simultaneously.

Here's the legal landscape as of April 2026: federal law covers intimate image deepfakes. State laws cover everything from election interference to commercial fraud. And the case law is being written right now. If your practice touches media, employment, entertainment, family law, or criminal defense, deepfake issues are already in your pipeline, or they will be within the year.

Are deepfakes illegal? What federal law says

Short answer (last verified October 30, 2026): Some are. The federal TAKE IT DOWN Act (Public Law 119-12, May 19, 2025) makes it a crime to knowingly publish a nonconsensual intimate deepfake of an identifiable person: up to 2 years in prison for adults, 3 for minors. It also requires covered platforms to remove reported images within 48 hours. Other deepfakes fall under state laws. We read the official statute text of 16 states on October 30, 2026 (table below); a state missing from the table may still have a deepfake law, so check the statute in each state that matters.

ItemWhat the Act saysSource
LawTools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks Act (TAKE IT DOWN Act), approved May 19, 2025 (S. 146)Public Law 119-12, Sec. 1
What it coversA “digital forgery” is an intimate visual depiction of an identifiable person created with software, machine learning, artificial intelligence or other technological means, including by altering an authentic depictionPublic Law 119-12, Sec. 2, 47 U.S.C. 223(h)
Criminal penaltyFine, imprisonment of not more than 2 years, or both for offenses involving adults; not more than 3 years for offenses involving minorsPublic Law 119-12, Sec. 2, 223(h)(4)
Platform dutyA covered platform must remove the depiction as soon as possible, but not later than 48 hours after a valid request, and make reasonable efforts to remove known identical copiesPublic Law 119-12, Sec. 3(a)(3)
Deadline to set up the processNot later than 1 year after enactment, so May 19, 2026 by our count, a covered platform must have a process for people to report such imagesPublic Law 119-12, Sec. 3(a)(1)(A)
Who enforces the platform dutyThe Federal Trade Commission: failure to reasonably comply is treated as a violation of a rule on unfair or deceptive acts or practicesPublic Law 119-12, Sec. 3(b)
Which platformsPublic-facing sites and apps that mainly host user-generated content, or that regularly publish or host such depictions; broadband providers, email and sites of mostly preselected content are excludedPublic Law 119-12, Sec. 4(3)

Last verified October 30, 2026, read in the Government Publishing Office text of the Act. This table covers the federal Act only; we did not re-read state statutes for this update. The Act sets out criminal and platform duties; it does not say a victim can sue under it.


The TAKE IT DOWN Act: Federal Criminal Liability

The TAKE IT DOWN Act (formally the Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks Act) creates federal criminal penalties for publishing non-consensual intimate images, including AI-generated deepfakes. Key provisions:

Criminal penalties: up to 2 years imprisonment for knowingly publishing non-consensual intimate deepfakes of an adult, and up to 3 years when the victim is a minor.

Platform obligations: social media platforms and websites must remove reported non-consensual intimate images within 48 hours of receiving a valid removal request. Platforms that fail to comply face FTC enforcement and potential fines.

Victim remedies: the Act itself does not create a private lawsuit. Its civil side is the removal duty, which the FTC enforces. Victims who want damages sue under state law and the other causes of action covered below, and those claims do not depend on a criminal prosecution.

The law's scope is narrow by design: it targets intimate images specifically, not all deepfakes. Political deepfakes, commercial deepfakes, and non-intimate fake content fall outside the federal statute and remain governed by state law, if any applies.

State Deepfake Laws: The Patchwork

State legislatures have enacted their own deepfake laws, creating a patchwork that practitioners need to navigate by jurisdiction. An earlier version of this guide gave state counts (43 states overall, with counts per category); we could not tie them to a primary source, so we removed them. Check the current statute in each state that matters to your matter.

Election deepfakes: a number of states prohibit AI-generated content depicting candidates in false scenarios within defined periods before elections. Texas, California, and Michigan have the broadest prohibitions. Penalties range from misdemeanor charges to civil fines up to $100,000. The constitutional challenges are coming: First Amendment arguments against these restrictions are pending in multiple circuits.

Non-consensual intimate deepfakes: many states have laws addressing this category, often predating the federal TAKE IT DOWN Act. State laws typically provide broader civil remedies and longer statutes of limitation than the federal statute. California and Illinois have the strongest victim protection frameworks.

Commercial deepfakes: some states address the use of AI-generated likenesses for commercial purposes without consent. These laws extend traditional right-of-publicity protections to AI-generated content. Tennessee's ELVIS Act (Ensuring Likeness Voice and Image Security Act) is the model statute, explicitly protecting vocal and visual likeness from AI replication.

Employment deepfakes: a few states address the use of deepfakes in employment contexts, fake reference videos, manipulated interview recordings, and fabricated credentials. Illinois's AI Video Interview Act requires disclosure when AI is used in hiring decisions, with deepfake provisions added in 2025.

State deepfake statutes we read in the official text

16 states, each row linked to the legislature or official code page we read on October 30, 2026. A state missing from this table may still have a deepfake law: it only means its official text was not readable that day.

StateStatuteWhat it coversIn force, as the source states itPrimary source
ArizonaA.R.S. 16-1023, digital impersonationCivil, election: a candidate or any citizen may sue; declaratory reliefThe section says it applies 14 days after its effective date; the page does not print the dateazleg.gov
ColoradoH.B. 24-1147, deepfakes in communications about candidatesElection: disclosure statement, fines, private action by candidatesTakes effect July 1, 2024; applies to communications distributed on or after that dateSigned bill (leg.colorado.gov)
FloridaFla. Stat. 836.13, altered sexual depictionsSexual images: criminal (third-degree felony)History lists ch. 2022-212 through ch. 2026-20leg.state.fl.us
FloridaFla. Stat. 106.145, use of artificial intelligence in political adsElection: disclaimer required; first-degree misdemeanorHistory lists ch. 2024-126leg.state.fl.us
IdahoIdaho Code 18-6606, disclosing explicit synthetic mediaSexual images: criminal (misdemeanor; felony for minors or repeat offenses)Added 2024 ch. 105; amended 2026 ch. 121legislature.idaho.gov
IdahoIdaho Code 67-6628A, FAIR Elections ActElection: civil, injunction and damages for the candidateAdded 2024 ch. 172legislature.idaho.gov
MassachusettsM.G.L. c. 265, s. 43A(b), criminal harassment, including digitized intimate imagesSexual images: criminalThe page does not print an effective datemalegislature.gov
MichiganMCL 168.932f, materially deceptive mediaElection: criminal (misdemeanor) and injunctive relief; 90-day windowAdded by 2023 Act 265, effective February 13, 2024legislature.mi.gov
MinnesotaMinn. Stat. 609.771, deep fakes to influence an electionElection: criminalNew in 2023 c 58 s 2; amended 2024 c 112revisor.mn.gov
MinnesotaMinn. Stat. 604.32, nonconsensual dissemination of an intimate deep fakeSexual images: civil action for damagesNew in 2023 c 58 s 1revisor.mn.gov
NevadaNRS 200.700, definition of computer-generated child sexual abuse materialSexual images of minors: the definition covers AI-created images indistinguishable from a minor; the offense sections are not covered hereThe page does not print an effective date for this sectionleg.state.nv.us
New MexicoH.B. 182 (2024), materially deceptive media in the Campaign Reporting ActElection: disclaimer in ads and a crime of distributingNot covered here; the row cites the enacted bill, not the current code sectionEnacted bill (nmlegis.gov)
North CarolinaN.C.G.S. 14-190.5A, disclosure of private imagesSexual images, including AI-created ones: criminal and civilThe page does not print an effective datencleg.gov
OhioR.C. 2917.211 and 2307.66, fabricated sexual imagesSexual images: criminal (fifth- or fourth-degree felony) and civilBoth pages: effective September 30, 2025 (H.B. 96)R.C. 2917.211; R.C. 2307.66
OregonS.B. 1571 (2024), synthetic media in campaign communicationsElection: civil penaltyThe enrolled bill declares an emergency; the exact date is not covered hereEnrolled bill (oregonlegislature.gov)
South CarolinaS.C. Code 16-15-330 and 16-15-332, digitally forged intimate imagesSexual images: criminal (felony when harm or profit is intended)Section 16-15-332 history: 2025 Act No. 37, effective May 12, 2025scstatehouse.gov
VirginiaVa. Code 18.2-386.2, unlawful dissemination of images, including modified images of a real personSexual images: criminal (Class 1 misdemeanor)History: 2014 c. 399; 2019 cc. 490, 515; 2024 c. 697law.lis.virginia.gov
WashingtonRCW 29B.35.010 (formerly 42.62.020), synthetic media in electionsElection: civil action by the candidate2023 c 360 s 2; recodified effective January 1, 2026app.leg.wa.gov
WashingtonRCW 9A.86.030, disclosing fabricated intimate imagesSexual images: criminal2024 c 88 s 7app.leg.wa.gov
WisconsinWis. Stat. 11.1303(2), synthetic media in political communicationsElection: disclosureHistory includes 2023 a. 123docs.legis.wisconsin.gov

The “in force” column repeats what each page prints (history note or effective date); a statute may have been amended since. Colorado, New Mexico and Oregon are cited from the enacted bill, not the codified section. This is not legal advice.

Civil Case Law: Where the Precedent Is Building

The case law is developing across multiple causes of action:

Defamation: courts are treating AI-generated deepfakes depicting real people in defamatory scenarios as actionable defamation. The publication element is satisfied by online distribution. The "of and concerning" element is met when the deepfake is recognizable as the plaintiff. The key question, whether deepfake creators can claim the content is "obviously fake" as a defense, is being litigated in multiple jurisdictions. Early rulings suggest that realistic deepfakes don't receive the same satirical protection as obvious parody.

Right of publicity: Tennessee, California, and New York courts have applied right-of-publicity claims to AI-generated likenesses. The emerging rule: using someone's likeness to create AI-generated content for commercial purposes without consent violates right of publicity regardless of whether the result is photorealistic. This extends to voice cloning: AI-replicated voices are treated as protectable aspects of personal identity.

Emotional distress: victims of non-consensual intimate deepfakes are bringing IIED claims in addition to statutory claims. Courts have consistently found that creating and distributing intimate deepfakes constitutes extreme and outrageous conduct meeting the IIED threshold. Damage awards in early cases range from $50,000 to $500,000.

Copyright: the intersection of deepfakes and copyright is murky. Using copyrighted photographs to generate deepfakes may constitute derivative works. Using a person's appearance (not copyrighted) to generate AI content doesn't trigger copyright. The training-data question, whether AI models trained on copyrighted images infringe by generating deepfakes, connects to the broader AI copyright litigation landscape.

Criminal Defense: Deepfakes as Evidence Challenges

Deepfake technology creates a novel evidence authentication challenge that criminal defense attorneys are already exploiting. The "deepfake defense", arguing that video or audio evidence could be AI-generated, has been raised in at least 47 criminal cases as of early 2026.

Authentication requirements: courts are tightening evidence authentication rules in response. The DOJ issued updated guidelines in 2025 requiring prosecutors to provide chain-of-custody documentation, metadata analysis, and forensic authentication for video and audio evidence challenged as potential deepfakes.

Expert testimony: deepfake detection experts are becoming a standard feature in cases where video evidence is contested. Forensic analysis tools (Microsoft Video Authenticator, Intel FakeCatcher, academic tools) can identify manipulation markers, but detection accuracy varies. Courts are applying Daubert analysis to deepfake detection testimony, and the standard for admissible detection methodology is still being established.

The defense attorney's dilemma: the deepfake defense raises legitimate questions about evidence reliability, but overuse risks undermining genuine evidence challenges. Courts are beginning to require attorneys raising the deepfake defense to proffer some basis for the challenge beyond theoretical possibility.

Practice Implications: What Every Attorney Needs to Know

For family law attorneys: deepfakes are appearing in custody disputes, divorce proceedings, and protective order hearings. Fabricated evidence showing a spouse in compromising situations, manufactured communications, and AI-manipulated audio recordings are all emerging issues. Counsel should proactively address deepfake authenticity in evidence motions.

For employment attorneys: AI-generated deepfakes in the workplace create hostile work environment claims (deepfakes of coworkers), discrimination claims (AI-generated content reflecting bias), and fraud claims (fabricated credentials or references). Advise employer clients to implement policies addressing AI-generated content and deepfake reporting procedures.

For IP attorneys: right-of-publicity claims for AI-generated likenesses are the fastest-growing area of deepfake litigation. Advise commercial clients that using any person's likeness in AI-generated content requires consent, even if the output isn't photorealistic. Voice cloning for commercial purposes requires the same consent framework.

For criminal defense attorneys: understand deepfake detection methodology well enough to challenge it effectively. The technology is new enough that forensic standards aren't settled: there's room for legitimate Daubert challenges to detection testimony. But raise the deepfake defense selectively, with specific technical basis, to maintain credibility with the court.

The Bottom Line: The deepfake legal landscape in 2026 spans federal criminal law (TAKE IT DOWN Act), state statutes covering election interference, intimate images, commercial use, and employment, plus rapidly developing case law in defamation, right of publicity, IIED, and evidence authentication. Every practice area is affected. Attorneys who understand the current framework will capture the wave of deepfake-related matters already building across civil and criminal dockets.

AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.

Next reads