How AI-Generated Fake Influencers Are Scamming Shoppers on TikTok

Fake influencers created by artificial intelligence are spreading across TikTok, Facebook, and Instagram to sell cheap mass-produced products, sometimes using fake Black personas in ways researchers call digital blackface.

The Verge found evidence of AI-generated accounts that invent entire life stories around products, claim to visit trade shows, and automatically reply to customer comments as if they were real people. One AI influencer called Aliyah has gathered 40,000 TikTok followers while promoting items through TikTok Shop, the platform's built-in shopping feature.

The deception goes further than just posting ads. These fake accounts create detailed stories about how products are made, pretend to attend industry events, and use automation to respond to comments in a way that looks like a real person is running the account. The money comes from dropshipping—a business model where the seller doesn't actually make or hold the products. The AI influencers promote these cheap items without telling anyone they're automated or have nothing to do with making the products.

How Big Is the Problem?

TikTok Shop has a fraud problem. The platform processes roughly 70 million products, according to available data, and many of those listings are fake or violate rules. Nicolas Waldmann, who leads safety efforts at TikTok Shop, says the platform uses its own AI systems to find and remove fraudulent content—which creates an odd situation where one AI fights another AI.

The numbers are staggering. TikTok has removed hundreds of thousands of fake sellers and fake product listings, but fraudsters create new ones faster than the platform can catch them. These scammers are using AI tools to build entire fake brands, fake product lines, and fake marketing videos across TikTok Shop and other sites like Amazon.

Jeremy Carrasco, a researcher who studies AI-generated videos and runs Riddance.ai (an organization that detects fake AI videos), explains that these fake influencers succeed because platforms don't verify who is really behind an account. The creator economy—where influencers earn money from sponsorships—is worth $11 billion per year. AI-generated influencers undercut real creators and trick shoppers at the same time.

What Are Regulators Doing About It?

The Federal Trade Commission has created rules to stop fake product reviews and reviews made by AI. However, these rules don't directly address synthetic influencers posting ads on social commerce features. The rules focus on review sections but miss the broader problem of fake personas promoting products.

It gets worse. Some scammers use AI to copy the faces of real creators. Charles Ray, an actual content creator, discovered his likeness stolen across multiple fake accounts claiming to run charities like animal rescues and churches—then selling products. These deepfake operations trick people by exploiting a creator's reputation and avoid having to disclose paid partnerships.

The technology varies in sophistication. Some fake influencers are just simple cartoon avatars with scripted responses. Others use more advanced natural language processing—a technology that lets AI understand and write conversational language—to reply to comments and generate videos that demonstrate products.

Why Is This Happening Now?

The creator economy was designed for human beings. Real influencers have limits: they sleep, they take breaks, they charge money for their time. AI accounts don't. They can post constantly, reply to every single comment instantly, and adjust their content to match what the algorithm rewards—all without the constraints that slow down human creators.

For brands trying to sell products, AI influencers look cheap and fast. But they carry a big risk: if people find out the account is fake, it damages the brand's reputation. And dropshipping makes it worse, because the products are often low quality and customer service is nonexistent.

Looking at how technology has disrupted markets before, this pattern is familiar. When digital advertising moved to automated buying—a system called programmatic advertising—it created huge opportunities for fraud. Scammers exploited automated systems designed to be efficient, not secure. We are seeing the same thing now, except this time the target is the person shopping, not the advertiser's budget.

The Architecture Problem

TikTok Shop lets you buy products without leaving the app. You see a video, tap to buy, and the product shows up on your doorstep. That's convenient for shoppers, but it's a vulnerability for platforms. Traditional online stores verify the seller, check that products are real, and build trust over time. TikTok Shop skips most of those checks—it just trusts that the person posting is real.

Algorithms also boost posts that get the most comments and replies. Fake influencers exploit this by using automation to reply to every single comment, giving their posts a burst of activity that makes them appear more popular than they actually are.

The fact that some AI influencers adopt Black personas for marketing purposes raises a concern worth flagging. When synthetic accounts use racial characteristics as a marketing tool, they treat identity as a commercial product while avoiding the actual lived experiences that shape authentic creators' perspectives. This extends the fraud beyond mere commerce deception into questions about representation and respect.

Platforms face a difficult choice going forward. Adding strict verification measures—proving you are who you claim to be—could protect shoppers but might also make it harder for legitimate creators to grow accounts. The technology for detecting synthetic content is improving, but so is the technology for creating it. For now, the fake influencers are likely winning the arms race.

The long-term outlook points toward a social commerce landscape where it becomes harder to know who or what is real. The financial incentives all favor sophisticated fake content. Detection requires human judgment that doesn't scale well across billions of posts. Without significant changes to how platforms verify creators, scams will probably become more common and harder to spot.