There is a woman on Instagram who has over 300,000 followers. She posts videos of herself visiting grocery stores, baking bread, and warning her audience about the dangers of processed food. She speaks with conviction. She looks real. She has been building her following for months.
She does not exist.
According to the New York Times, she is one of several fully AI-generated personas created specifically to sell a wellness supplement online. No disclosure. No fine print. Just a synthetic character with a carefully built backstory, a large audience, and a product to push.
This is not an isolated case. It is a business model.
How it actually works
The setup is simpler than you would expect. A marketer generates a fictional character using AI tools, builds out a personality and a visual identity, starts posting content, and grows a following. Once the audience is large enough and trusts the persona, the character starts recommending products. Supplements, detox powders, fibre drinks, wellness tonics.
The same operator can run multiple characters simultaneously. An Amish woman. A monk. A cluster of nearly identical fitness personalities. All different faces. All promoting the same product. All fake.
Because these characters are AI-generated, there are no contracts to negotiate, no human creators to manage, no risk of a public scandal. The content runs continuously. The characters never have a bad day. According to the New York Times, one operator described this approach as a "game changer" and said every part of his business was being "AI-ified."
The products being sold have not necessarily been clinically tested. In at least one reported case, the operator behind the campaign said he planned to conduct a study to see whether the product actually worked. The study had not yet been done. The product was already on sale.
It goes further than fake personas
Invented characters are only part of the picture. A separate and more troubling tactic involves deepfaking real people.
According to Full Fact, academics and credentialed health professionals have had their likenesses used without consent in supplement ads on TikTok and Instagram. Videos showed fabricated versions of real researchers discussing health topics they never commented on, directing viewers toward products they had never endorsed. In one documented case, a university professor discovered over a dozen videos on TikTok featuring a synthetic version of himself making false health claims. He filed a complaint. The platform took six weeks to remove them.
Fast Company reported that scammers are exploiting popular wellness trends on short-form video platforms by creating AI-generated influencers with elaborate fake backstories, building large followings, and then selling health products through fabricated personal testimonials. Some accounts accumulated millions of likes before the accounts were removed. Several were structured as networks, using identical content formats and promoting the same products across multiple synthetic identities.
The Better Business Bureau has also flagged a surge in AI-assisted health scams in 2026, including deepfake videos falsely showing well-known public figures endorsing weight loss products. One example generated over 170 consumer complaints with victims spending hundreds of dollars each.
The research backs up why this works
A study published in the British Journal of Psychology in February 2026 found that people consistently overestimate their ability to identify AI-generated faces. The technology has reached a point where synthetic faces are difficult to distinguish from real ones even for people who believe they can spot the difference. Marketers are aware of this. The entire business model depends on it.
Timothy Caulfield from the Health Law Institute at the University of Alberta told the New York Times that AI tools allow companies to cheaply test dozens of different synthetic spokespeople until they find one that resonates with an audience. There is no meaningful cost to failure. They just generate another character and try again.
Regulators are starting to move
New York became the first US state to pass a law specifically requiring disclosure when synthetic performers are used in advertising. The law, signed in December 2025 and taking effect in June 2026, defines a synthetic performer as any AI-generated character intended to create the impression of a real human being. Violations carry fines starting at one thousand dollars per incident.
The FTC has also signalled a shift toward active enforcement rather than guidance, sending warning letters to companies for potential violations and expanding its focus to include AI-generated endorsements in sensitive categories including supplements and health claims.
Whether enforcement will keep pace with how fast the technology is moving is a different question entirely.
The bottom line
Influencer marketing was already an industry built on manufactured trust. AI has now made that manufacturing cheap, scalable, and almost impossible to detect with the naked eye.
The characters recommending products to you on social media may have large followings, convincing personalities, and detailed backstories. They may look and sound completely real. That does not mean they are. And the products they are selling you are real enough, as is the money you spend on them.
The next time a wellness influencer tells you they have discovered something that changed their health, it is worth asking one basic question. Does this person actually exist?
Increasingly, the answer is no.
