Fake AI Influencers Are Selling Cheap Knockoffs on TikTok

Computer-generated people are appearing across social media — TikTok, Facebook, Instagram — to promote cheap products that they don't actually make or own. These fake influencers are powered by artificial intelligence, and The Verge found that some of them are pretending to be Black, raising questions about how AI is being used to deceive both shoppers and creators.

One AI influencer named Aliyah built up 40,000 followers on TikTok by promoting products through the app's shopping feature. She doesn't exist. Her posts tell fake stories about making products, claim she's at trade shows, and seem to answer customer questions — but it's all automated. She's programmed to create a convincing illusion of a real person running a real business.

The business model behind these fake influencers is called dropshipping. Here's how it works: the AI account promotes a product, customers buy it through TikTok, but the actual product comes from a warehouse somewhere else — often cheap mass-produced goods. The AI influencer takes a cut without ever touching the product or caring whether it's any good.

How Big Is This Problem?

TikTok says it's processing violations across 70 million products, according to available data. The company has removed hundreds of thousands of fake sellers and fraudulent listings. Nicolas Waldmann, who leads governance for TikTok Shop, confirmed that the platform uses its own AI to detect these fake accounts — essentially fighting AI with AI.

The problem is speed. Creating fake influencers and fake listings through AI happens faster than people can remove them by hand. Fraudsters can now generate entire fake brand identities — the stories, the products, the promotional videos — all automatically.

Jeremy Carrasco, a researcher who studies AI-generated media and runs an organization called Riddance.ai that focuses on detecting AI videos, says these operations work because platforms don't verify who creators really are before letting them sell. The broader creator economy — the ecosystem of influencers and brands working together — is worth 11 billion dollars a year, which makes it an attractive target for fraud.

What About the Law?

The Federal Trade Commission introduced rules against fake and AI-generated reviews, but those rules don't directly address AI-generated people pretending to be influencers. The law is catching up slowly.

Scammers have also created AI versions of real, established creators. One creator named Charles Ray discovered fake accounts using his face, claiming to run animal rescues and farms and churches, all as cover to sell products. These deepfake operations — where AI copies someone's appearance — take advantage of the trust people have in real creators.

Some of these fake accounts use simple AI avatars with pre-written responses. Others are more sophisticated and use advanced language systems to write natural-sounding replies to comments and even generate videos. The quality varies, which gives platforms a way to catch them — if they invest in tools to do so.

Why This Matters for Real Influencers

The bigger context is worth considering. Tech platforms are increasingly automating their advertising systems to save money and reduce human involvement. When that happens, AI-generated personas can slip through more easily because fewer people are watching.

This creates real pressure on human creators. A fake AI influencer can post new content 24 hours a day without getting tired. It never asks for a raise. It responds to every comment instantly. Real influencers can't compete with that kind of availability. For brands, the appeal is obvious: AI influencers are cheap and scalable.

But there's a catch. If people discover the influencer is fake, the brand's reputation takes a hit. And the dropshipping model means the fake influencer is promoting products it has never tested and doesn't vouch for. If customers receive poor quality items, the brand suffers.

We have seen this pattern before, when programmatic advertising — software that automatically buys and places ads — made it easier to commit ad fraud at scale. The difference now is that the deception is aimed at individual shoppers rather than at businesses. It's more personal and harder to detect.

The Structural Problem

TikTok's shopping feature lives right inside the app. You watch a video, you like what you see, you buy it — all without leaving TikTok. That's convenient for shoppers but risky for verification. On traditional online stores, you can check seller history, product reviews, and customer feedback before buying. TikTok Shop collapses those verification steps into a single question: do you trust the creator.

AI-generated personas are especially dangerous in this environment because they can mimic real creator behavior. If the platform's algorithm favors accounts that get lots of comments and replies, a fake influencer can game that system by automatically responding to everything, boosting its visibility beyond what human creators achieve.

The bigger question here is about how AI is being used to copy identity itself. Some of these fake influencers adopt racial and ethnic characteristics to seem authentic. When synthetic personas borrow racial identity purely for marketing — to sell more dropshipped goods — they're treating identity as a prop. That raises concerns that go beyond fraud and touch on how AI represents people.

What's Next

The technical battle between those creating fake influencers and those trying to detect them will intensify. Both sides are improving their tools. But platforms face a genuine tension: adding more verification steps to catch fakes could also make it harder for legitimate creators to reach audiences.

The direction this is heading is toward a world where it becomes much harder to know whether you're really talking to a person or a program on social media. That's a problem that detection software alone can't solve. Eventually, platforms may need to add ways for creators to prove their identity — similar to how banks verify account holders. But that would slow down the frictionless shopping experience that makes TikTok Shop attractive in the first place.

For now, the economic incentives strongly favor sophisticated fake influencers. Real detection requires human judgment, and platforms can't hire enough people to review billions of posts. That's the gap where fraud thrives.