What Are Deepfakes: Fake Shriver AI Ads Cost Victims $500

Maria Shriver did not sell you a memory supplement. She wants that said plainly, because somewhere out there, a video is circulating that says otherwise — her face, her voice (or something close enough to fool you), telling you to buy a pill for Alzheimer's disease. She never recorded it. She never approved it. And if you're asking what are deepfakes and why they suddenly look this convincing, her situation is the answer.

TL;DR: Scammers are using AI to fake celebrity endorsements — including Maria Shriver's — for Alzheimer's supplements that don't work, and the video looks real enough to fool a tired person scrolling at 11pm.

Here's the part that should actually worry you: this isn't just Shriver. Bill Gates, Anderson Cooper, Bruce Willis — all of them have had their faces stitched onto fake videos pushing a "honey-based remedy" for Alzheimer's. None of them made those videos. Somebody with editing software and an AI tool did, and then bought ads to put it in front of you while you were scrolling, worried about your own memory or your parent's. This is deepfake technology working exactly as designed: convincing, cheap to produce, and aimed straight at your trust.

What are deepfakes, and why does deepfake technology work so well on health ads

A deepfake is a fake video or audio clip made by AI that swaps a real person's face, body, or voice onto footage they never actually filmed. Think of it as a very convincing digital puppet wearing someone else's face. The tech behind it — sometimes called synthetic media — relies on machine learning and artificial intelligence models trained on huge datasets of real footage, and it has gotten good enough that odd blinking or mismatched lips, the old tells that once made deepfake detection easy, don't always show up anymore. A June 2025 study found people trust AI-generated "doctor" avatars more when the avatar looks older — which tells you the scam artists aren't just faking faces, they're faking the specific kind of face your brain is wired to trust.

That's the real trick. It's not really about fooling your eyes. It's about hijacking a shortcut your brain uses every day: if someone I trust says this is safe, I don't need to check it myself. Psychologists call this authority bias — and it's usually a good thing, it's how society functions without everyone re-verifying every claim from scratch. But when the "trusted person" is a fabricated video wearing a stolen face, that shortcut becomes the exact door the scam walks through. This article is part of a series — start with How To Spot A Deepfake.

Is the Maria Shriver deepfake ads security threat a real, ongoing scam right now

Yes. Shriver posted a public statement after AI-made ads began circulating that falsely showed her endorsing Alzheimer's supplements. She's one of many public figures targeted — not a one-off, but part of a pattern investigators are now tracking across multiple countries and platforms as a growing security concern.

Inside the fake Alzheimer's supplement scam targeting real people

Let's get specific, because the vague version of this story ("AI is scary") doesn't help you at all. The pattern, according to reporting from NBC Chicago, goes like this: a video ad shows up in your social feed, a familiar face is talking with total confidence about a supplement — often something honey-based — that supposedly slows or reverses memory loss. There's a link. You click. You buy.

Then the second scam starts. According to reporting on victims, people have been hit with unauthorized recurring credit card charges they never agreed to, customer service lines that go nowhere, and refund requests that get denied outright. One victim reported being charged over $500 within just ten days of clicking a single ad. Richmond Side documented a case involving a man deceived by fake videos of Steve Martin promoting a supplement called MemoPryl — a product with no real backing, wrapped in a stolen celebrity endorsement. The video and voice work in these deepfakes is polished enough that fake detection by eye alone is no longer reliable.

Shriver has noted that these ads are especially cruel because scammers are deliberately targeting millions of people who are terrified of Alzheimer's disease — preying on both their fear and their hope for a cure. — Maria Shriver, The Women's Alzheimer's Movement

That's the whole game, isn't it? Fear plus hope is the most profitable emotional combination in advertising, real or fake. Someone watching a parent slip into forgetfulness will click almost anything that promises relief — especially if the promise comes wrapped in a face they've trusted for decades.

What does deepfake detection look like on a fake supplement ad video

Usually a short video clip, 15-30 seconds, showing a celebrity "explaining" a supplement breakthrough, followed by a link to a sales page with urgent language ("limited stock," "as seen on TV") and a subscription checkout — the part where the real financial damage starts. Fake videos like these rely on borrowed data, not new evidence, to seem believable.

Why deepfakes have turned trust into a fraud and security target

Here's where this stops being a Maria Shriver story and becomes an everyone story. Research published in Frontiers in Communication makes the case that endorsement advertising has always worked less on the actual product claim and more on who's making it. This kind of data science research into synthetic media matters because you don't fact-check the honey cure — you fact-check the person. And AI has figured out exactly how to counterfeit that second thing — the identity, the authority, the decades of built-up credibility — without needing to touch the truth of what's being sold at all. Previously in this series: Deepfake Legislation No Law Stops Ai From Training On You.

That's a strange, slightly disturbing shift. It used to be that a fake product needed a fake claim. Now a fake product just needs a real face and enough video polish to pass a casual glance. Some analysts estimate deepfake celebrity-baiting ads now account for more than half of all reported synthetic media fraud cases in 2026 — meaning this isn't a fringe scam anymore, it's arguably the dominant one. The content behind these deepfakes keeps improving, which is exactly why detection needs to shift from "look for glitches" to "verify the source."

The one thing worth checking before you buy anything from a celebrity ad

If you've ever half-wondered whether a video is really who it claims to be, that's not paranoia — that's the exact question this whole mess forces us to start asking. So here's one concrete, doable thing: before you buy anything from a celebrity health endorsement you saw in an ad, go directly to that celebrity's own verified website or their official social media account and search their name plus the product. Real endorsements show up in more than one place — press releases, interviews, their own posts, and consistent data across every channel. A fabricated one exists only inside that one ad, nowhere else, because nobody real ever agreed to say it.

Shriver's own team took this exact step — putting a plain, dated statement on The Women's Alzheimer's Movement site denying any connection to these products. That's the pattern to look for: if a celebrity is real, their actual channels will confirm or deny it directly. If they're silent everywhere except the ad itself, that silence is the answer.

What are deepfakes doing in Maria Shriver's fake supplement ads

In Shriver's case, the point is not to win an argument about AI regulation. It's to stop a person worried about Alzheimer's from mistaking a paid social-media ad for a real recommendation. The FTC already has authority over deceptive advertising, and reporting from STAT News notes that AI-generated medical misinformation videos are squarely inside its jurisdiction, and that information security agencies are watching this space closely, even if enforcement has not caught up to the volume of fake ads flooding the web. Up next: How To Spot A Deepfake 1 School Photo Is All It Takes.

The familiar warning signs — odd blinking, strange hands, glitchy jewelry — may help, but they are not enough to rely on. In these supplement ads, the useful check is simpler: pause before paying, leave the ad, and look for the endorsement on Shriver's own channels or those of any celebrity shown. A real endorsement can be confirmed outside the ad; a fabricated one cannot.

So no, you're probably not going to out-stare an AI video until it glitches. What you can do is refuse to let a familiar face do your thinking for you — especially when the ask is your credit card and the topic is something as personal as Alzheimer's. Shriver didn't sell that supplement. Neither did Bill Gates, Anderson Cooper, or Steve Martin. The only thing that actually sold it was your trust, quietly borrowed without asking. Deepfakes, deep fakes, whatever you call the technology behind them, work the same way every time: they don't need to fool your eyes forever, just long enough to get your card number.

Frequently Asked Questions