TL;DR: Are deepfakes illegal? Making a deepfake is not illegal on its own. What the law punishes is how it is used: sexually explicit fakes made without consent, deceptive election content, scams and impersonation, and commercial use of someone's face or voice without permission. In the U.S., the federal TAKE IT DOWN Act now makes publishing nonconsensual intimate deepfakes a crime and forces platforms to remove them within 48 hours, and nearly every state has added its own rules. In the EU, deepfakes must be labeled from August 2026. If you create AI video, avatars, or voice content, the safe path is simple: get consent, label synthetic media, and never fake a real person in a way that could deceive or harm them.
Deepfake laws are changing faster than almost any other area of law. Three years ago, most states had nothing on the books. Today, 48 states regulate sexually explicit deepfakes, 33 regulate political deepfakes, and Congress has passed its first federal deepfake statute.
That matters if you are a creator using AI video tools, a marketer running ads with digital avatars, a business worried about voice-clone fraud, or someone who has found a fake image of yourself online. This guide covers the federal law, the main state laws (with a comparison table), the rules in the EU and UK, and what is actually legal when you make synthetic media.
This article is general information, not legal advice. Deepfake rules depend on your state, the content, and how it is shared, so speak with a licensed attorney about a specific situation.
Key Takeaways
- No blanket ban. Deepfakes are legal to create in most contexts. Liability attaches to specific harms: intimate imagery, election deception, fraud, defamation, and unauthorized commercial use of a likeness.
- Federal law: the TAKE IT DOWN Act (signed May 19, 2025) criminalizes publishing nonconsensual intimate deepfakes. Since May 19, 2026, platforms must remove them within 48 hours of a valid request.
- Pending, not passed: the NO FAKES Act (a federal likeness right) and the DEFIANCE Act (a federal civil lawsuit for victims) are still bills as of October 2026.
- States lead: every state except Ohio and New Mexico now addresses sexually explicit deepfakes, and 33 states regulate election deepfakes (Ballotpedia, August 2026).
- Labeling is becoming mandatory: EU AI Act Article 50 requires deepfakes to be disclosed from August 2, 2026, and New York now requires ads to disclose AI-generated "synthetic performers."
Are deepfakes illegal? The short answer
A deepfake is AI-generated or AI-altered audio, video, or images that realistically show a person saying or doing something they never said or did. The technology itself is lawful. Film studios use it for de-aging, brands use licensed digital doubles, and creators use AI avatars of themselves every day.
What makes a deepfake illegal is its content and purpose. Across U.S. federal and state law, five categories carry real legal risk:
- Nonconsensual intimate imagery (NCII). Sexually explicit fakes of a real person made or shared without consent. This is the most heavily regulated category, and it is criminal almost everywhere.
- Election deception. Fakes of candidates or election officials distributed close to an election to mislead voters, usually without a disclaimer.
- Fraud and impersonation. Voice clones used to trick a bank, an employee, or a family member; fake executive video calls; robocalls using AI voices.
- Unauthorized commercial use. Using a real person's face or voice to sell a product without permission. This falls under the right of publicity and, increasingly, specific "digital replica" laws.
- Defamation and harassment. A fake that damages someone's reputation or is used to threaten them can support traditional claims like defamation, false light, or intentional infliction of emotional distress.
If your content falls outside all five, you are generally in lawful territory, though platform rules and labeling laws may still apply.
Federal deepfake laws in the United States
There is still no single federal law covering all deepfakes. Instead, a handful of targeted rules apply.
The TAKE IT DOWN Act
The TAKE IT DOWN Act was signed on May 19, 2025, and is the first federal statute aimed squarely at deepfakes. It does two things:
- Criminal ban. It is a federal crime to knowingly publish nonconsensual intimate images, including AI-generated "digital forgeries," of an identifiable person. Threatening to publish them is also a crime. Penalties are higher when the person depicted is a minor.
- Platform takedown duty. Since May 19, 2026, covered platforms (social networks, image and video sites, messaging and gaming services that host user content) must provide a removal process and take down reported images, plus known identical copies, within 48 hours of a valid request.
The FTC enforces the platform side. It sent compliance letters to more than a dozen major companies before the deadline, including Meta, TikTok, X, Reddit, and Snapchat, and launched TakeItDown.ftc.gov so victims can report platforms that fail to act. A violation is treated as breaking an FTC rule, which carries civil penalties of up to $53,088 per violation.
The FCC ruling on AI voices in robocalls
In February 2024, the FCC unanimously ruled that AI-generated voices count as "artificial or prerecorded" voices under the Telephone Consumer Protection Act. In practice, any robocall that uses a cloned or synthetic voice needs the recipient's prior consent and must identify the caller. The ruling followed the fake "Biden" robocalls sent to New Hampshire primary voters. For a deeper look at voice-specific rules, see our guide to whether AI voice cloning is legal.
Federal bills still pending
Two bills get cited as if they were law. They are not, at least not yet:
| Bill | What it would do | Status (October 2026) |
|---|---|---|
| NO FAKES Act (S. 4591 / H.R. 8915) | Create a federal right to control digital replicas of your voice and likeness, with notice-and-takedown for platforms | Cleared Senate Judiciary unanimously in June 2026; a fast-track floor vote was blocked on September 30, 2026 over parody and satire concerns |
| DEFIANCE Act | Give victims of sexually explicit deepfakes a federal civil claim with statutory damages | Passed the Senate unanimously in January 2026; awaiting a House vote |
Until either passes, a victim's civil lawsuit runs through state law.
Deepfake laws by state
States moved first and still do most of the work. According to Ballotpedia's August 2026 report, states enacted 58 deepfake bills in the first part of 2026 alone, most of them on sexually explicit content.
Here is how the most-searched states compare:
| State | Sexually explicit deepfakes | Election deepfakes | Likeness and voice rights |
|---|---|---|---|
| California | Civil claim (Civ. Code 1708.86) and criminal penalties for distributing intimate fakes | AB 2839 and AB 2655 (2024) were both struck down in federal court in Kohls v. Bonta | Strong right of publicity (Civ. Code 3344); AB 2602 limits digital-replica clauses in performer contracts; AB 1836 requires estate consent for replicas of deceased performers (both effective 2025) |
| Texas | Penal Code 21.165 covers produced or distributed sexually explicit deepfake videos and, since September 1, 2025, images too; threats to share are also an offense | Election Code 255.004 bars deceptive deepfake videos within 30 days of an election | Common-law and statutory publicity rights |
| Minnesota | Civil and criminal remedies for nonconsensual sexual deepfakes | Stat. 609.771 bans deceptive election deepfakes within 90 days of an election; a challenge was rejected at the preliminary stage by the Eighth Circuit in 2026, though litigation continues | Common-law publicity rights |
| Tennessee | Covered under state intimate-image laws | Varies; check the current election code | ELVIS Act (2024) protects a person's voice, including AI simulations, and reaches tools whose primary purpose is cloning someone without authorization |
| New York | Civil Rights Law 52-c lets victims sue over sexually explicit fakes; Penal Law covers unlawful dissemination | Requires disclosure on materially deceptive election media | Post-mortem digital replica rights (Civil Rights Law 50-f); synthetic performer disclosure required in ads since June 9, 2026 |
Two patterns stand out:
- Election laws face First Amendment pushback. California's two election deepfake laws were struck down, one on free speech grounds and one because Section 230 preempted it. Most surviving election laws exempt content that carries a clear disclaimer, which is why labeling is the safest habit.
- Intimate-image laws are expanding, not shrinking. Texas widened its law from videos to images in 2025. Almost every state now treats nonconsensual sexual deepfakes as a crime, a civil wrong, or both.
State statutes change every session. To check the current text of a specific law and how courts have read it, AI case law search pulls the decisions with links to the full opinions.
Understand your rights before you sign
Review contracts, understand legal notices, and get clear answers about everyday legal issues — grounded in real law with citations you can check.
Get legal answersDeepfake laws outside the U.S.
If your content reaches European audiences, two regimes matter.
European Union: AI Act Article 50. From August 2, 2026, anyone who deploys an AI system to create a deepfake must disclose that the content is artificially generated or manipulated. Providers of generative tools must also mark outputs in a machine-readable way (systems already on the market before August 2 have until December 2, 2026). There is a lighter touch for evidently artistic, satirical, or fictional work: disclosure is still required, but it can be done in a way that does not ruin the work. Fines reach EUR 15 million or 3% of global turnover. The 2026 "Digital Omnibus" delayed the AI Act's high-risk rules, but it did not delay these transparency duties.
United Kingdom. Sharing intimate deepfakes without consent has been an offense under the Online Safety Act since January 2024. Since February 6, 2026, it is also a crime in England and Wales to create, or ask someone to create, a sexually explicit deepfake of an adult without consent, even if it is never shared.
AI likeness rights: when can you use someone's face or voice?
For creators and marketers, the practical question is less "is this a deepfake?" and more "do I have the right to use this person?"
The right of publicity protects a person's name, image, likeness, and in many states voice against unauthorized commercial use. Most states recognize it, by statute, common law, or both. Using an AI replica of a celebrity to endorse your product is the classic violation, and the new digital replica laws in California, New York, and Tennessee make the AI angle explicit.
Here is a working checklist:
- Your own likeness: fine. Training an avatar on your own face and voice is the safest use case.
- Employees, actors, or influencers: get written consent that names the specific uses (ads, platforms, duration, languages) and whether a digital replica can be created. A vague "all media, forever" clause may be unenforceable for performers in California.
- Public figures in commentary: parody, satire, news, and criticism get First Amendment protection. A clearly labeled satirical sketch is very different from a realistic fake endorsement.
- Deceased people: many states protect publicity rights after death (California for 70 years). You may need the estate's consent.
- Private individuals: avoid realistic fakes of private people without consent. This is where harassment, defamation, and NCII exposure concentrate.
Ownership is a separate question. Even if you have the right to use a likeness, who owns the finished AI video depends on copyright rules for machine output. Our guide on who owns AI-generated content covers that side.
How to use AI video and avatars legally
If you create synthetic media for marketing or content, these habits keep you on the right side of almost every law above:
- Get documented consent from anyone whose face or voice you replicate, and keep it on file.
- Label synthetic content. A visible "AI-generated" note satisfies EU Article 50, the disclaimer exemptions in most state election laws, and platform policies on TikTok, YouTube, and Meta.
- Disclose AI performers in ads. New York requires it, and the FTC treats undisclosed fake endorsements as deceptive. See our breakdown of FTC disclosure rules for influencers.
- Never fake real people in sexual content, even as a joke. This is criminal under federal law and in nearly every state.
- Stay away from election content near voting days unless it is clearly labeled.
- Read your AI tool's terms. Most generators prohibit impersonation and require you to warrant you have rights to uploaded faces and voices. Breaking those terms can also cost you your account.
- Review talent and agency contracts for digital replica clauses before you sign. AI document review can flag vague likeness grants and missing consent language.
What to do if you are the victim of a deepfake
If someone has made a fake of you, act quickly:
- Preserve evidence. Screenshot the content, the URL, the account, and the date before it is removed.
- Request removal. For intimate images, use the platform's TAKE IT DOWN process; the 48-hour clock starts with a valid request. If the platform ignores you, report it at TakeItDown.ftc.gov. StopNCII.org can also help block re-uploads.
- Report crimes. Intimate deepfakes, extortion, and fraud should go to local police, and to the FBI's IC3 portal for online fraud.
- Consider a civil claim. Depending on your state, claims may include a deepfake-specific statute, right of publicity, defamation, or intentional infliction of emotional distress. Statutes of limitations can be short.
- Talk to a lawyer who handles privacy, defamation, or internet law in your state.
Check the deepfake law that applies to you
Because deepfake laws differ so much between states and change every legislative session, the hard part is finding the current rule for your exact situation. LegesGPT's legal AI chatbot answers questions like "is this AI ad legal in Texas?" or "can I sue over a deepfake in New York?" with citations to the statutes and cases behind each answer, so you can verify the source before you rely on it or bring it to an attorney.
Frequently Asked Questions
Are deepfakes illegal?
Not in general. Creating a deepfake is legal in most contexts, such as parody, film effects, or an avatar of yourself. Deepfakes become illegal when they are sexually explicit and made without consent, used to deceive voters near an election, used for fraud or impersonation, or used to sell a product with someone's likeness without permission.
What is the TAKE IT DOWN Act?
The TAKE IT DOWN Act is a federal law signed on May 19, 2025. It makes it a crime to knowingly publish nonconsensual intimate images of an identifiable person, including AI-generated ones, and since May 19, 2026 it requires covered platforms to remove reported images within 48 hours of a valid request. The FTC enforces the platform obligations.
Is there a federal law against deepfakes?
Only for specific harms. The TAKE IT DOWN Act covers nonconsensual intimate deepfakes, and an FCC ruling requires consent for robocalls that use AI voices. Broader proposals such as the NO FAKES Act, which would create a federal likeness right, and the DEFIANCE Act, which would let victims sue, had not become law as of October 2026.
Which states have deepfake laws?
Nearly all of them. According to Ballotpedia's August 2026 report, every state except Ohio and New Mexico has a law addressing sexually explicit deepfakes, and 33 states regulate deepfakes in political communications. The details, penalties, and exceptions differ widely from state to state.
Can I use AI to make a video of a celebrity?
You can for clearly labeled parody, satire, commentary, or news, which receive First Amendment protection. You generally cannot use a celebrity's AI likeness or voice to advertise or endorse a product without permission, because that violates the right of publicity and, in states like California, New York, and Tennessee, specific digital replica laws.
Do I have to label AI-generated videos?
Increasingly, yes. The EU AI Act requires deepfakes to be disclosed from August 2, 2026, New York requires ads to disclose AI-generated synthetic performers, most state election laws require a disclaimer, and major platforms like YouTube, TikTok, and Meta require labels for realistic synthetic content.
What can I do if someone made a deepfake of me?
Save evidence first, including screenshots, URLs, and dates. Ask the platform to remove it, using its TAKE IT DOWN process for intimate images, and report non-compliant platforms to the FTC. Report extortion or fraud to the police, and talk to a lawyer about civil claims such as defamation, right of publicity, or your state's deepfake statute.
Is it illegal to make a deepfake for personal use?
For most content, private creation is not illegal. Sexually explicit deepfakes are the exception: in England and Wales it has been a crime since February 2026 to create one of an adult without consent even if it is never shared, and several U.S. states also penalize production, not just distribution.
