AI-generated doctors and deepfake supplement ads: how to check

You usually cannot tell by looking, so the safer habit is to check the source, not the face. A doctor, a celebrity or a "customer" in a supplement ad can now be produced or altered with AI, and the result can look and sound convincing for the few seconds an ad runs. This guide explains what regulators have said about AI impersonation, the clues worth looking for, and a short verification routine that works even when the video looks perfect.

One caution up front: none of these clues proves a video is fake, and a clean video does not prove it is real. What you can check is whether the page gives you a real, findable person and company behind the claims. Our review method treats that as a transparency question and scores it.

What is a deepfake supplement ad?

It is an ad in which the face, voice or both of a real or invented person are generated or altered by software. The person appears to say something they never said, or a person who does not exist appears as a doctor, a customer or a "researcher". The FTC has described the wider problem in plain terms: in February 2024 it warned that emerging technology, including AI-generated deepfakes, threatens to turbocharge impersonation fraud, and it said that fraudsters are using AI tools to impersonate individuals at a much wider scale.

The same FTC announcement says it proposed extending its impersonation rule to cover the impersonation of individuals, and that it was seeking comment on whether firms that provide AI tools they know are being used to harm consumers through impersonation should be covered too. That was a proposal at the time of the announcement, so check the FTC site for the current status before you rely on it. What matters for a shopper is the direction: regulators treat a synthetic person who vouches for a product as a serious problem, not a harmless production trick.

Does the FTC allow AI-generated people in ads?

Not without limits, and the rules are more specific than many people expect. The FTC's rule on consumer reviews and testimonials says its ban covers testimonials that misrepresent that they are by someone who does not exist. The agency's announcement of the final rule gives "AI-generated fake reviews" as an example, and it prohibits businesses from creating or selling such fake or false testimonials, and from buying or spreading them when the business knew or should have known they were fake.

The FTC's question-and-answer page on the rule adds a useful nuance. It says the rule has no blanket prohibition on the use of AI-generated avatars in marketing, and that an avatar would only be prohibited as a testimonial if the underlying testimonial were fake or false. It also says the use of avatars could be deceptive under the FTC Act, and that it would violate the rule to use a celebrity avatar, without the celebrity's permission, to speak favorably about a product if reasonable consumers would think the celebrity actually gave a testimonial.

In other words, an animated narrator is not automatically illegal. A narrator who claims a real experience nobody had, or a famous face that never agreed to appear, is the problem.

What red flags suggest an ad uses an AI doctor?

Look for a mismatch between how much authority the person projects and how little you can verify. These are clues, not proof:

• A "doctor" with no full name, no license, no clinic and no publications you can look up.
• A polished talking head that never appears anywhere else: no conference talk, no clinic website, no interview, no author page.
• A voice that sounds slightly flat, with mouth movements that do not quite match the words, or hands and jewelry that shift oddly. These artifacts are getting rarer, so their absence means little.
• Stock-style lab imagery: white coats, microscopes and molecular graphics. The FTC's health products guidance gives this exact scenario as an example. It says images of doctors in white lab coats, microscopes and stacks of journals give an impression of scientific legitimacy and can convey an implied claim that the product has been clinically proven.
• A story told in first person by someone whose identity is only a first name, a city and a job title.

Some of the flags in our own reviews are of this kind. The The Brain Song review notes that the disclaimer admits avatars and people connected to the creators, and scores it 4.0 out of 10 on transparency. The BioDentex review records an unnamed doctor endorsement with no institution or credentials to check. We report what the pages say and do not claim to know how any video was made.

What about AI-made images and AI-written content?

Sales pages use AI in quieter ways than a talking deepfake. Our Quantum Brainwave Protocol review flags a main product graphic with visibly garbled, misspelled cover text, consistent with an AI image generator rather than a designed product photo. Our KetoFlow review notes before-and-after style images that are AI-made. And the herbal remedies ebook review credits the page for disclosing that the book is AI-generated, then points out that there is no sign of a human expert reviewing the doses or safety warnings.

The lesson is the same in each case. An AI-made image is not a fraud by itself. It becomes a problem when it is presented as evidence: a picture of a "result", a "customer", an "expert" or a "product" that a reader is meant to treat as real.

How do I check whether a spokesperson is real?

Use this routine, which takes about five minutes:

1. Get the name. No full name means there is nothing to check. Treat the endorsement as absent.
2. Search the name in quotes. Add "supplement", "clinic" or "license". A real physician usually has a footprint that predates the product.
3. Check where the video lives. Does it appear on a verified account of the person or institution, or only on the seller's page and in ads?
4. Run a reverse image search on a screenshot. Google explains how to search with an image using Google Lens. Results can include websites with the same image or a similar one, which helps you find the same face under different names. No match is not proof of authenticity.
5. Find the company. Look for a legal name, street address and support email in the footer or terms page.
6. Judge the claim, not the face. Ask what the person says and whether a study, not a testimonial, supports it.

The FTC's health products guidance makes a related point about experts. Its example of an infomercial "Doctor" who was not licensed to practice medicine, had not reviewed the science and was paid with an officer position at the company concludes that the ad was likely to be deceptive, partly because the qualifications were overstated and partly because the connection to the company was not one consumers would expect. Real credentials and disclosed relationships are what an endorsement needs. A convincing face is not.

Why does a paid or invented expert still matter if the ingredients are fine?

Because the expert is usually what makes a weak claim feel strong. A page can list ingredients with real research behind them and still lean on a doctor, a logo or a famous name to suggest more than the label supports. Even an openly paid expert is not independent. Our Pep Tonic review credits the page for disclosing that a named physician is paid by the company, and adds that a paid physician is not a study.

If you cannot find a person, a study and a company you can check, then what you are looking at is marketing, and it should be weighed as marketing. If you are researching a specific health problem, the category pages and the full review list show how each offer handles this.

What should I do if I think an ad used a fake or cloned person?

You have three practical options. Report the ad on the platform where you saw it, using its report tool. Report it to the FTC at ReportFraud.ftc.gov, which the FTC lists on its announcements as the place to report fraud. And if you already paid, use the steps in our guides to refunds through ClickBank or refunds through Digistore24, and ask your card issuer about a dispute if the seller does not respond.

The FDA also runs a health fraud page for products that claim to prevent or treat diseases without proof. It warns that health fraud scams waste money and can delay proper diagnosis and treatment. If you have a condition, talk to your doctor or pharmacist before starting or stopping anything, and never stop a prescribed medication because of an ad.

FAQ

Can I trust a doctor in a supplement ad if the video looks real?

Not on the video alone. Trust what you can verify: a full name, a license or institution you can look up, and a claim backed by a study you can open. Even a real doctor may be paid, and a paid endorsement is not independent evidence.

How can I tell if a celebrity or doctor video was made with AI?

Often you cannot, and the visual clues are getting weaker. Check the source instead. If the person really backed the product, you would expect to find the statement on their own verified channels or in news coverage, not only in ads and on the seller's page.

Are AI avatars in ads illegal?

Not automatically. According to the FTC's Q&A on its reviews rule, there is no blanket ban on AI avatars, but a fake or false testimonial is prohibited, and using a celebrity avatar without permission to praise a product is a violation when consumers would think the celebrity really endorsed it.

Where do I report a fake endorsement?

Use the report tool on the platform where you saw the ad, and file a report with the FTC at ReportFraud.ftc.gov. Keep screenshots and the page address.