Celebrity deepfake scams: how to check suspicious endorsements

A suspicious endorsement may show a famous investor promoting an investment, an actor recommending a product, or a CEO fronting a giveaway. The face and voice may look convincing. That appearance is a claim to verify, not proof.

What the FBI actually warns about

The FBI's public warning (IC3, PSA241203) is blunt: criminals exploit generative AI to commit fraud at a larger scale and make it more believable. The FBI notes that generative AI can reduce the time and effort criminals spend deceiving a target, and can even correct human errors that might otherwise warn you something is wrong.

Among the tactics the FBI describes, two bear directly on endorsement scams: criminals use AI-generated video to create believable depictions of public figures that bolster their fraud schemes, and AI-generated audio to impersonate well-known figures to elicit payments. The FBI also notes that AI-generated text drives investment and other confidence schemes, and that AI-generated images produce believable profile photos and identification documents. Creating or distributing synthetic content is not inherently illegal, the FBI adds, but it can facilitate fraud and extortion.

Why a celebrity's face or voice is no longer proof

The urge to trust a familiar face is the reflex these scams rent. A convincing likeness of a public figure does not tell you the endorsement is real — and, just as much, an awkward or imperfect clip does not tell you it is fake. Treat the identity in the video as a claim to be verified, not as evidence in itself. The question is never "does this look like them?" but "can I confirm this from a channel that person actually controls?"

A five-step check for a suspicious endorsement

Work outward from the content instead of staring at the face:

1. Open the person's or brand's official channel yourself. Type the known website or verified handle directly — don't follow the link in the video or ad. Check whether the person or company confirms the campaign there. If you cannot find confirmation, do not rely on the clip alone.
2. Read the domain, not the logo. Compare the landing-page domain with the official domain you opened yourself. Extra words, unusual suffixes, or misspellings are reasons to stop and verify before entering information or paying.
3. Interrogate the money ask. Write down exactly what payment is requested, who would receive it, and why. Do not pay until the person or brand confirms the request through an independently opened official channel. The FBI warning specifically links AI-generated impersonation to eliciting payments and investment fraud schemes.
4. Find the original clip. Search a distinctive phrase from the clip, or reverse-image-search a frame, to look for an earlier or original publication. If you find one, compare its image, audio, context, and destination link. If you do not, the clip remains unverified.
5. Get a second, independent signal. If you have the file, run it through a detector for one more data point — but keep reading on how to weight that result.

Where a detector helps — and where it doesn't

A detector gives you a probabilistic risk signal, not a verdict. DeepFakeCheck can analyze an image, video, audio clip, or block of text and return the likelihood that it was AI-generated or manipulated; the first check needs no signup and there is a limited free allowance. As one input among several, that is genuinely useful.

Be clear about the limits, though. Automated detectors produce both false positives (flagging authentic media) and false negatives (missing a manipulated clip). A score — high or low — cannot on its own prove that an endorsement is genuine, that a person actually said something, or that a piece of media is authentic. Use the number to raise or lower your suspicion, and let the channel checks above make the real decision.

If money is already involved

If you've sent funds — or nearly did — treat it as suspected fraud, not embarrassment. Stop any further payments, contact your bank or card issuer immediately, and preserve everything: the video, the URL, the payment details, and any messages. The FBI directs victims of financial fraud to report to its Internet Crime Complaint Center at www.ic3.gov.

The FBI warns that generative AI can reduce the time and effort needed to deceive a target and make a scheme more believable. Your defense isn't a sharper eye for artifacts — it's the habit of leaving the content and verifying the source.

Sources

FBI Internet Crime Complaint Center (IC3), public service announcement, "Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud" — https://www.ic3.gov/PSA/2024/PSA241203