Making a deepfake is not automatically illegal in the United States, but the moment you share one, use one to deceive, or create certain categories of content, you can run into federal crimes, state crimes, and civil lawsuits all at once. Whether making deepfakes is illegal depends on what the deepfake shows, who it depicts, how it’s used, and where you are. The clearest no-go zones are sexual deepfakes of real people, any sexual imagery of minors, deepfakes used to defraud, and deceptive political deepfakes near an election. Everything else sits on a spectrum of civil risk.
When a Deepfake Becomes a Federal Crime
The TAKE IT DOWN Act, signed into law in 2025, is the first federal statute that directly criminalizes non-consensual intimate imagery, and it explicitly covers AI-generated forgeries alongside authentic photos and videos. The crime is knowingly publishing an intimate visual depiction of an identifiable person without consent through interstate commerce.1Congress.gov. S.146 – TAKE IT DOWN Act 119th Congress (2025-2026)
Publication is the trigger at the federal level. If you generate a sexual deepfake of a real person and post it anywhere online, you have potentially committed a federal offense even if you never charged anyone money, never named the subject, and took it down quickly. The law also forces platforms to remove reported images within 48 hours, enforced by the Federal Trade Commission.1Congress.gov. S.146 – TAKE IT DOWN Act 119th Congress (2025-2026)
State Laws on Sexual Deepfakes
State law reaches further than the federal statute, and it was reaching further long before Congress acted. By mid-2025, roughly 45 states criminalized the creation or distribution of sexually explicit deepfakes without the depicted person’s consent.
Penalties vary. Some states treat distribution as a misdemeanor with fines in the range of $2,500 to $5,000. Others charge it as a felony with meaningful prison time. Prosecutors generally must prove the creator or distributor knew the subject hadn’t consented, and some states additionally require an intent to harass, intimidate, or harm.
The important point for anyone asking whether making a deepfake is illegal: several state statutes criminalize the creation or possession of a non-consensual sexual deepfake even if you never share it. Federal law targets publication; a chunk of state law targets the act of making it. If you build a sexual deepfake of a real person in your home and keep it on your own hard drive, you may already have broken state law.
Deepfakes of Minors
AI-generated sexual imagery of a minor is treated as child sexual abuse material under existing federal law. Federal obscenity statutes already cover computer-generated depictions of minors engaged in sexually explicit conduct, so prosecutors do not have to prove a real child was involved. Penalties match those for CSAM involving actual children: long prison terms, mandatory sex offender registration, no statute of limitations.
Congress is working to tighten this further. The ENFORCE Act, which passed the Senate unanimously in December 2025, would clarify that producing AI-modified CSAM carries the same penalties regardless of whether the offender intended to distribute it, and would align penalties across the different federal statutes that prosecutors might use.2Office of Senator John Cornyn. Cornyn, Blumenthal, Lee, Kennedy Bill to Prosecute AI-Generated CSAM Passes Senate Unanimously The bill is awaiting House action as of early 2026. Even without it, federal prosecutors already have the tools to charge AI-generated CSAM.
Using a Deepfake to Defraud or Impersonate
If a deepfake is a tool in a scheme to take money or steal an identity, the underlying federal fraud statutes supply the charges, and the ceilings are high.
The federal wire fraud statute covers any scheme to defraud carried out through electronic communications and carries up to 20 years in prison. When the fraud affects a financial institution, the ceiling rises to 30 years and fines up to $1,000,000.3Office of the Law Revision Counsel. 18 US Code 1343 – Fraud by Wire, Radio, or Television Voice cloning used to impersonate executives and trigger wire transfers has become the common corporate pattern, with some incidents producing losses in the tens of millions of dollars. The deepfake is the method; the crime is the fraud.
Deepfakes used to bypass biometric security or to fraudulently obtain personal information can also bring federal identity fraud charges. Using someone else’s identifying information to commit a federal crime or a state felony carries up to 15 years in prison, rising to 20 years when connected to drug trafficking or violent crime.4Office of the Law Revision Counsel. 18 US Code 1028 – Fraud and Related Activity in Connection with Identification Documents, Authentication Features, and Information
Deepfakes in Political Campaigns
Political deepfakes carry their own layer of state and federal rules. By the end of 2024, roughly 20 states had enacted laws regulating deepfakes in elections, and the number continues to grow. The specifics vary. Texas treats distribution of a political deepfake within 30 days of an election as a misdemeanor. California prohibits deceptive AI-generated election content distributed within 120 days of an election.5The First Amendment Encyclopedia. Political Deepfakes and Elections
Most state election statutes take one of two forms. The more common requires a clear disclosure label stating that the material was created with artificial intelligence. A smaller group bans deceptive political deepfakes outright, label or no label. Violations usually carry civil penalties, and affected candidates can seek court orders to stop distribution.
At the federal level, the Federal Election Commission adopted an interpretive rule in September 2024 confirming that the existing ban on fraudulent misrepresentation of campaign authority under 52 U.S.C. ยง 30124 is technology-neutral. Using a deepfake to impersonate a candidate or falsely claim to speak on their behalf violates the statute the same way any other method of fraud would.6Federal Election Commission. Commission Approves Notification of Disposition, Interpretive Rule on Artificial Intelligence in Campaign Ads
Civil Lawsuits You Could Face
Even where no criminal statute reaches a particular deepfake, the person depicted can often sue. Two civil theories do most of the work.
Right of publicity protects a person’s control over the commercial use of their name, voice, and likeness. It’s governed by state law, recognized in most states by statute or common law. A deepfake that uses someone’s face or voice to sell a product, endorse a brand, or generate revenue without permission is a textbook violation. Statutory damages in states that specify them typically range from $750 to $10,000, with actual and punitive damages pushing totals higher when financial harm can be proven.
Defamation applies when a deepfake falsely portrays someone doing something illegal, immoral, or otherwise damaging to reputation. A fake video of a person committing a crime they didn’t commit meets the elements of libel: a false statement of fact, published to others, causing reputational harm. Public figures must show the creator knew the content was false or acted with reckless disregard. Private individuals generally only have to show negligence.
The First Amendment is a weaker shield here than people assume. Courts have held that constitutional protection does not cover fraud, commercial misappropriation, or content made with actual malice. Parody and satire are protected, but only when a reasonable viewer would recognize the content as non-literal commentary rather than a genuine depiction. A photorealistic deepfake labeled as satire somewhere in a caption is a weak candidate for that defense.
Copyright Risk from Source Material
Deepfakes can also infringe copyright independent of any claim by the person depicted. Building a deepfake from copyrighted source material, such as a copyrighted photograph, film clip, or recording, can violate the copyright holder’s exclusive rights to reproduce the work and to make derivatives from it.
The U.S. Copyright Office in a 2025 report concluded that assembling a training dataset from copyrighted works “clearly implicates the right of reproduction,” and that if a resulting AI model retains substantial protectable expression from its training data, even the model weights could constitute infringement. The main defense is fair use, analyzed case by case on purpose, nature, amount used, and market effect.7United States Copyright Office. Copyright and Artificial Intelligence, Part 3 – Generative AI Training Recent court decisions have gone both ways, with courts rejecting fair use where the AI tool directly competed with the market for the original and where licensing markets for training data already existed. The law is still shifting.
What’s Coming Next
Two pending federal bills would substantially widen the liability picture for deepfake creators.
The DEFIANCE Act would give victims of explicit deepfake imagery a federal civil cause of action with liquidated damages up to $150,000, rising to $250,000 when connected to sexual assault, stalking, or harassment. It would carry a 10-year statute of limitations and allow victims to proceed under pseudonyms. It passed the Senate unanimously and was awaiting House consideration as of early 2026.8Problem Solvers Caucus. Problem Solvers Caucus Endorses DEFIANCE Act to Allow Victims of Non-Consensual Deepfakes to Sue Perpetrators
The NO FAKES Act would create the first federal intellectual property right in a person’s voice and likeness, replacing the current state-by-state patchwork. It would prohibit the non-consensual use of digital replicas in sound recordings and audiovisual works, require written consent describing the intended use and a limited term, and set up a mandatory takedown process. It carves out protections for news, commentary, criticism, satire, parody, and documentary work.9Congress.gov. S.1367 – 119th Congress (2025-2026) – NO FAKES Act of 2025
Neither has been signed into law. But the unanimous Senate votes point in one direction: the window in which making a deepfake is only loosely regulated is closing. If you are considering making one of a real person, assume the law covering it next year will be tighter than the law covering it today.