Are AI Celebrity Voices Legal? A Plain-English Guide for Creators

Are AI celebrity voices legal? A plain-English guide to parody, right of publicity, the NO FAKES Act and election rules, plus a checklist for safe use.

Written byTrump AI Voice Team
Published
Reading time
10 min read
Abstract audio waveform passing through a balanced scale and a shield shape, in gold and violet tones on a dark background
Table of Contents

You can now type a sentence and hear it read back in a voice that sounds a lot like a president, a pop star, or a cartoon sponge. That raises an obvious question: is that even legal?

The honest answer is it depends on what you do with it. The technology itself isn't illegal in most places. What gets people in trouble is how they use it: to deceive, to sell, or to scam.

Here's a plain-English guide to where the lines are as of November 2026. It isn't legal advice. If you're planning anything commercial, talk to a lawyer.

The short answer

Usually OKRiskyLikely illegal
Obvious parody and satire, labeled as AIAds, sponsorships, or anything that implies endorsementScams and fraud using someone's voice
Private jokes among friendsRealistic fake statements about real eventsRobocalls with AI voices made without consent
Commentary and criticism that's clearly fictionalUsing voices of deceased performers commerciallyElection disinformation designed to deceive voters
Fan content with no money involvedSelling a tool built mainly to clone specific peopleSexual or defamatory content about real people

The more your use looks like "this is clearly a joke," the safer you are. The more it looks like "this person actually said this" or "this person backs this product," the more legal risk you take on.

Most people assume a voice is protected by copyright. It mostly isn't.

In July 2025, a federal court in New York ruled in Lehrman and Sage v. Lovo, a case brought by two voice actors against an AI voice company. As the law firm Skadden summarized it, the court found that copyright protects specific recordings, not the qualities of a voice itself. Most of the federal trademark and copyright claims were dismissed.

But the voice actors' state-law claims went forward, including claims under New York's right of publicity law. That's the key point. In the US, your voice is protected mainly by a patchwork of state right-of-publicity laws, which stop people from using your identity for commercial purposes without permission.

That idea is older than AI. In Midler v. Ford Motor Co. (1988), Ford hired a singer to imitate Bette Midler in a car commercial after Midler declined, and a federal appeals court sided with Midler. In Waits v. Frito-Lay (1992), Tom Waits won a case over a snack ad that used a sound-alike of his gravelly voice. Both cases were about advertising, and that's still where the risk is highest.

When it's usually fine: parody and satire

In the US, the First Amendment gives strong protection to parody and satire of public figures. In Hustler Magazine v. Falwell (1988), the Supreme Court held that public figures can't win damages for emotional distress over parody that no reasonable person would take as a statement of fact.

Courts have applied that thinking to AI, too. In August 2025, a federal court ruled that California's 2024 law restricting deceptive election deepfakes violated the First Amendment, in a case brought by a creator of AI parody videos. According to Columbia University's Global Freedom of Expression database, the court found the law swept in protected speech like satire and parody.

What makes parody safe is that it's obvious. A clip of a famous voice wishing your cat a happy birthday isn't fooling anyone. A realistic clip of the same voice "admitting" to a crime is a different story.

What helps parody stay parody:

  • Absurd, clearly fictional content
  • A label like "AI parody" in the caption or audio
  • No connection to real, current events presented as fact
  • No money changing hands for endorsements

When it gets risky

1. Ads and endorsements

This is the clearest danger zone. Using a celebrity's voice, real or AI, to sell something implies they endorse it. That's what Midler and Waits were about.

New laws make this even riskier:

  • Tennessee's ELVIS Act took effect on July 1, 2024. It added voice explicitly to the state's right-of-publicity law, and also targets distributing tools whose primary purpose is reproducing a specific person's voice or likeness without authorization.
  • California passed AB 2602 and AB 1836 in September 2024. They limit contract terms that let companies replace performers with digital replicas, and extend post-mortem publicity rights to cover digital replicas of deceased performers.
  • New York signed two laws on December 11, 2025. One requires ads to disclose when they use AI-generated "synthetic performers." The other strengthens protection against unauthorized digital replicas of deceased performers.

Bottom line: Don't use a celebrity's voice, or an obvious imitation of it, in any ad, sponsored post, or product promotion without a license.

2. Impersonation and scams

Using someone's voice to make people believe they're actually talking to that person is where civil risk becomes criminal risk.

This isn't hypothetical. In May 2025, the FBI warned that criminals were using AI-generated voice messages to impersonate senior US officials, mainly to get into the personal accounts of government officials and their contacts.

Phone calls have their own rules. In February 2024, the FCC ruled that AI-generated voices count as "artificial" voices under the Telephone Consumer Protection Act, so robocalls using them need the same prior consent as any prerecorded call. The ruling followed AI robocalls that imitated President Biden's voice before New Hampshire's January 2024 primary.

That case is instructive. The political consultant behind the calls, Steve Kramer, was acquitted of state voter suppression and candidate impersonation charges by a New Hampshire jury in June 2025, according to NHPR. But the FCC had separately proposed a $6 million fine against him, and the phone company that carried the calls agreed to pay $1 million in a settlement. Being found not guilty in one court doesn't make you safe from regulators.

3. Realistic fake statements

A realistic clip of a real person "saying" something about real events can create problems even if you meant it as a joke: defamation, false-light claims, platform bans, and, during elections, a growing list of state laws.

The test to ask yourself: if this clip were cut out and shared without your caption, would people believe it? If yes, rewrite it or don't post it.

4. Deceased celebrities

Many states protect a person's identity for years after death. California and New York have both strengthened those rights for digital replicas. If you're making anything commercial with the voice of a performer who has died, assume you need permission from the estate.

What's happening at the federal level

There's still no single federal law on AI voice clones in the US. The main proposal is the NO FAKES Act. Its 2026 version (S.4591 in the Senate and H.R.8915 in the House) was introduced on May 20, 2026. It would create a federal right in a person's voice and visual likeness, plus a notice-and-takedown system for unauthorized digital replicas. The Senate Judiciary Committee approved it unanimously in June 2026, according to Deadline.

As of late October 2026, it had not passed either chamber. The current Congress ends on January 3, 2027, so if it doesn't pass by then, it would have to be reintroduced. Earlier versions of the bill included exclusions for uses like news, commentary, criticism, satire, and parody. Check the final text if it passes.

Outside the US

The European Union

Since August 2, 2026, the EU AI Act's transparency rules have required people who publish deepfakes, including realistic AI-generated audio, to disclose that the content is artificial. For works that are evidently artistic, creative, satirical, or fictional, the disclosure duty is lighter: you still disclose, but in a way that doesn't spoil the work. Lawyers reading the European Commission's guidance note that this lighter treatment doesn't cover identity fraud. We explain who has to do what in our guide to the EU AI Act content labeling rules.

Denmark

On October 8, 2026, Denmark's government brought a bill to parliament that would ban sharing realistic deepfakes of someone's face or voice without their consent. According to Euronews, the bill keeps satire, parody, caricature, and social criticism allowed, and the government hopes it will take effect at the start of 2027.

Platform rules matter too

Even when something is legal, a platform can still take it down. YouTube has required creators to disclose realistic altered or synthetic content since March 2024, including generated voices narrating a video. Most major platforms have similar labeling rules. Breaking them can get your content removed or your account restricted.

A safety checklist for AI celebrity voices

Before you post, run through this list:

  1. Is it obviously a joke? Would a stranger instantly know it's not real?
  2. Is it labeled? "AI parody" in the caption, on screen, or in the audio.
  3. Is it free of commercial use? No ads, no sponsorships, no implied endorsements.
  4. Is it free of real news? No fake statements about real events, elections, or policies.
  5. Is it honest? No pretending to be a real person on a call, in a message, or in a DM.
  6. Is it decent? Nothing sexual, hateful, or defamatory about a real person.
  7. Does it follow platform rules? Check the disclosure settings when you upload.
  8. Is the person private? If it's a voice clone of a private individual, get their consent. Ordinary people have even more reason to object than celebrities do.

If you answer "no" to any of the first six, change the content before you post.

Using AI celebrity voice tools responsibly

Reputable tools make the rules part of the product. The Trump AI voice generator at Trump AI Voice, for example, says its voices are for entertainment and parody only, states it isn't affiliated with Donald Trump, and asks users not to use the audio to deceive people, impersonate anyone, or create content that could be mistaken for real statements. Its AI celebrity voice generator applies the same rules to its other voices and asks users to label audio as AI-generated. You can make a clip right on Trump AI Voice, which makes it easy to keep everything in the parody lane: write something absurd, generate it, label it, share it.

The bottom line

AI celebrity voices are a legitimate tool for comedy, parody, and fan content. The law in most places protects obvious jokes about public figures. What it doesn't protect is using someone's voice to sell, to deceive, or to defraud.

Keep it silly, keep it labeled, keep money out of it, and you'll stay on the right side of the line almost everywhere.

This article is a general explanation, not legal advice. Laws differ by state and country and are changing quickly.

Sources

Create with Trump AI Voice

Turn your text into a Trump AI voice

Create expressive speech with the Trump AI voice generator.

Try Trump AI Voice

Written by

Trump AI Voice Team

All articles