Celebrity deepfakes, and what to do about them.

What is a celebrity deepfake?

A celebrity deepfake is an AI-generated video, image, or audio clip that copies a famous person's face or voice to make them appear to say or do something they never did. Most are made for one of two reasons: attention, or money. The money kind is the dangerous one, because a familiar face makes people drop their guard.

The scale is new. A single fake can be cloned into hundreds of versions across video, image, and ad platforms within days, and each one points a real person's audience at something the person never approved.

Celebrity deepfake scams and fake endorsements

The most common commercial abuse is the fake endorsement. An AI version of a well-known actor, host, or athlete appears to promote a product, a crypto scheme, a giveaway, or a supplement. The person had nothing to do with it. Recent cases have hit household names across TV, film, and sport, and the pattern is always the same: borrow the trust the celebrity built, convert it into clicks and payments.

These are not harmless. They defraud the public and they damage the person whose face is used, who then spends time and money cleaning up a mess they did not make.

Are celebrity deepfakes illegal?

Usually, when they are commercial. Using a celebrity's likeness or voice to sell or promote without permission runs into right-of-publicity law in most US states, and into newer AI-specific rules. Deceased celebrities are covered too: states like California require estate consent for a digital replica of a late performer. Sexual deepfakes are separately illegal under federal law. For the statute detail see our guide on deepfake laws.

The claim almost always exists. The hard part is finding every copy and proving it, before it moves or disappears.

What to do about a celebrity deepfake

Whether you are the talent, the agency, or the estate, the order is the same. Find all of it first: one reported fake usually means many, so search across video, image, audio and ad platforms, not just the source.

Then preserve the evidence. Capture each fake with a timestamped, tamper-evident record, because scam accounts get taken down or edited fast and a screenshot is easy to dispute. How talicense preserves evidence →

Report and take down next: platforms have impersonation and right-of-publicity routes, and a clean record makes them faster. Then escalate the commercial ones, the paid scam campaigns and repeat infringers, where the documented record is what a lawyer acts on.

talicense finds unauthorized AI uses of a person's likeness across platforms and documents each one so it holds up, then the rights holder decides every move. See how PROTECT works →