Let’s be honest about what these videos look like. Then imagine them. A smiling woman in her forties, makeup perfect, kitchen bright and clean, holds up a bottle of dark green powder and says, “This is what we look like on Day 1 of taking beetroot every day. This is Day 50.” She seems energetic, happy, trustworthy. Another woman, perhaps with silver hair and a white lab coat, says, “This is Day 77, girl,” as if she is talking to an old friend across a kitchen table. They look exactly like the wellness influencers you scroll past a dozen times a day on TikTok, Instagram, and Facebook. But none of them are real. Their faces were generated by artificial intelligence. Their voices may be synthetic, too. And their entire job is to sell health supplements, many of them unproven, unregulated, or even linked to recall. The Times reviewed hundreds of these advertisements and found that A.I.-video tools have created a bizarre, underground marketplace of health disomation. A global network of anonymous companies, content farms, and digital hustlers is using fake people to sell real pills to real Americans. The ad says “beethroot” is a miracle; the next says moringga cures a disease; the one after that claims a tea reverses kidney failure. The people in these ads are never ill. They never get old. They neverhe public health. They are never accountable. They are only there to make a profit off vulnerability, hope, and sometimes desperation.
You can understand what that means if you spend time with someone like Pamela Wundrow. Pamela is 71 years old, a retired truck driver, and she carries more chronic weight in her body than many people realize. She has an autoimmune disease that causes joint problems, inflammation, and fatigue every single day. “I’ve always had the joint problems, and the inflammation, and the fatigue, so I basically wanted something that was going to help me with all that,” she says in a voice that sounds tired but not defeated. She remembers exactly why she trusted the ads for a supplement called Roseabella that featured a attractive herbal formula. She says she “does foraging,” a naturalist at heart, drawn to anything plant-based. Moringga seemed harmless. Maybe the plant could do something. She placed an order for the popular online brand, paid her money, and began taking it every morning. Soon, her old health issues got worse. Her complaint was not a standard side effect. It was a deep, nagging gut-feeling that something inside had been messes up. “When you already have a health problem before this and then you take this and all of a sudden you got worse, was this part of it? I don’t know. I can’t say, and I don’t know how they could prove it either.” Then in February, the U.S.F.A. and C.D.C. issued a public recall for Rosabella’s moringa supplement. Salmonella had been found. People have been sick. “It says fluousness all over the recall. But Pamela never tested positive for that. She no longer doesn’t experience the worst of it. She only knows the last few months. But she got paid. And all she has is uncertainty.
The truth is that the people behind these products rarely have names that anyone can know. Ambrosia Brands, the company behind the Rosabella scandal, did not respond to calls from The Times, not once, not multiple times, not in any way. What the company did leave behind is not worth the phone call refunds but a trail in Better Business Bureau records. Hundreds of written complaints over three years, each containing months of nothing but noise. And yet the company kept scaling, kept selling, kept releasing new formulas, kept advertising on social media. The algorithm kept it propped up it was profitable, and the casualties were women like Pamela. The manipulation of trust is not accidental again. To make viewers feel safe, the ads include familiar imagery: American flags, white suburban homes, clean kitchens, and smiling grandparents. The avatars are engineered——are repeated, too. Some wear chef coats yan medicinal robes. Their fake offices havecasionally diplomas on the walls, somehow a Korean university that doesn’t exist, invented to impress. In other ads, the same digital face is a doctor in one video, a nurse in another, and an “Eastern expert” in the third. Chinese medicine and Western practice are tier together to create an illusion or authority. It is easy to laugh if you see the “AI” artifact of a missing hand or a stiff elbow, but many of these videos are almost perfect. Pamela said watching one of them: “I mean I believe in Chinese medicine, and I can see how this would hook people because she looks real. I have seen others where you can tell it’s A.I. To me this does not look like A.I.’s.” That is the most dangerous line in the whole story; when the visual no longer leaks the truth, humans not detect what must have.
It gets even worse when you look at the claims some of these videos are willing to make. On television, the same advertisement could trigger an FDA warning for a prescription drug, but on social media, it thrives. In one set of ads, a long-time actress posed as if she were a scientist and says, “We gave 100 people with stage 3 kidney disease a bottle of Sculptique and told them” — the caption implies that ell herbal pill save them from dialysis. That title is enough to kill. Fake well being influencers like this say that a supplement can substitute for a blood pressure drug, a statin, or insulin. Instead of going to a doctor, people turn to a $29 online bottle. There are no federal rules stopping the use of synthetic people in commercials. No federal law says a voice must be human. On Facebook and Instagram, Meta says it labels AI-generated content from third-party models, but the labels are often missing to the company at all. The videos still live for when they reported; Facebook removed them after they were exposed. TikTok also says its paid policy bans accounts that move that misinformation. But the rules are reactive, not preventive. By the time a platform counts the issue, the ad has already circulated millions and sold thousands.
Who is behind this machine? The answer is global and from its mailing list. A Chinese agency that specialists in AI ads has produced more than 1,200 videos per day in just the past year for its un 500 in trading. For about $10, the company creates a talking, breathing, truly bad American-shaped avatar. It sells inside TikTok Shop and Amazon; supplement powders and capsules are uploaded into fake product categories. Vivi Yi, who runs one of the agencies, refuses to show specific supplement ads because they violate contract with brands. She’s considerate of the danger: “Health products need extra caution on TikTok,” she says, but her agency doesn’t wave the line if the money interrupts. The company works in a gray area: they use US flags, culturally familiar, American voices, but they operate from an office in Guangzhou. The feeds is a geometry of disossociation: a Chinese made fake American avatar in a Korean female coathelated Indian algorithm, selling a Chinese-manufactured “herbal cure” to a retired truck driver in Middle America. No one is sure why a bottle again.
Yet technology is moving much faster than the law. Any person with a computer and an internet connection can now create a “realistic” medical spokesperson in less than ten minutes. Once these videos are generated, they are shared rapidly into Facebook groups, grief groups, geo-caching. The platforms have no human reviewer; they rely on moderation systems that miss most adversariality and allow the same bot across 200 accounts. Many online supplement brands are anonymous LLCs hidden behind mailboxes and shell companies. They absolutely do not care if you fall sick. The creators in China have no FDA contract. The U.S. creators who license them are not much better. Digital mimicry is now indistinguishable for many people, especially older voters, veterans, and those with already undefined chronic conditions. When they are on screens, they feel real. The fake influencer answers the questions with confidence and a motherly smile. They say “you want real truth,” but the trace is that they are alternative to the anti-real: no history of illness, no body, no intestine, no headache, no blood essay. They have never seen a clinic waiting room, never had a needle in their arm, never watched a relative die of cancer, never heard a doctor tell you there is none they can do. They can be perfect because they were built to be perfect—fake.
That is what creates the biggest harm. It is not simply that the pills sometimes carry sewage bacteria. It is that this entire world of pseudo-science—which seems trustworthy and intimate, created for a single moment in your bedroom—destroys the possibility of true trust. American health care has: misinformation filled, online communities. People who cannot find a doctor, cannot afford one, cannot get an appointment for more months, turn to social media. They find a face that says they understand. Good by the platform that just keeps pushing. No one checks the real credentials. No one inspects the office. The bottle with a “Moring” label arrives, costs fifty dollars, and goes in every body. The officials remove one or two products, or the company rebrands and returns as “Ne.Moringa” or a “Golden-LeafCalm” under some new website. The creator in another country makes a nervous video, starts selling and repeat. The fraud is structural, not individual.
We are now at a moment where “synthetic” has become a generic technology, not a footnote for creators. There is no federal law in the United States that specifically says misinformation cannot be placed in synthetic people’s mouths. A.I. influencers do not need to disclose themselves on their label anymore they have bewildered the same standard infiltration. Big Tech platforms measure engagement, not truth. They remove content after 50 pbs reported, after the sales is already closed. We protect a fake voice of a fake doctor more than we protect a real user who calls in distress. Maybe the only meaningful repuls solution is not a single platform policy or a single company. It will be a set of legal, political, and human guardrails: laws that make synthetic people unable to make health claims; mechanisms to automatically detect decompensation in lip movements and sound waves; if the product can make a medical claim, you proof to human being behind the brand; and one stant legal authority to track down anonymous owner. More importantly, the AI commercially should be forced to “You are a robot, and not allowed to advertise to a vulnerable person in crisis.”
But before we get there, consider the paras of Pamela again. She lives with invisible pain. She does not want to be scammed; we all want real. She wanted a simple herb to make her body less tired; that is not silly or a middle-age need. It is a human need in a system that flies them. When she looks at the face generated by A.I., she smiles a little and says “she looks.” She is not mistaken. The technology has become so perfect that it has outsmarted our perception. That is not a moral weakness in her; that is a moral failure in the system.
The only way to step out of that is to insist that trust itself must come from transparency. Explain to users: “If this bot put this, please. This does not mean the product is safe.” The ad should say it. The platform should say it. The government should say it. And the hashtag should say it, only perhaps these avatars. We have spent decades to create perfect influencers. We need to now spend all that energy marking flawed humans real. Perhaps the repair to grace the iron will be the deepest invisible line?
Pamela still isn’t convinced if she can explain. She wants an answer, but won’t. She wants a cure, but cannot find it. Maybe she won’t know how many times her trusting, her 71-year older mind was used by a product who flushed blood through them and somewhere else. But nothing will hurt her until the next ad: an AI face, functionally familiar, says “Moringa is natural.” She doesn’t click. She pauses, no doctor, no origins, no friend, no real trustworthy. Better than what? she says maybe, and lets scroll.
That is the next real counsel: a fake influencer might have taken awareness, and be careful. A “real” can remove hundreds of fake humans, but no, not if they are echoed, never. We have to face a human reality: trust is hard to rebuild once it has been stolen by a machine. The goal, then, is not merely make AI-free content. It is to make a whole industry of these wellness con men illegal. It is to say to Pamela: you deserve something better than after.
We need to remember who that fake “you” is really means: for profit, not for health. That should be enough to make you immediately honest. Because a real person cannot be refounded, but a viral advert can. We must to build a clear identity: true advertising, true health care, true and no actual both. The avatars can be used for clever entertainment, but never for fear, pain, or illness. We are all their doctors soon; if we let them, they will use our weakness. Let’s not allow.