AIで「偽の黒人」を生成してSheinの粗悪品を販売する詐欺師たち AI grifters are creating fake Black people to sell Shein junk

A troubling trend has emerged on TikTok's rapidly expanding live commerce platform: AI-generated fake Black personas being used to hawk cheap dropshipped goods, raising urgent questions about racial exploitation, consumer deception, and platform accountability.

The scheme works like this. Bad actors use AI image and video generation tools to create convincing avatars — often light-skinned Black women, sometimes dressed in relatable styles like country-western gear — and build social media accounts around them. These fake personas then sell products sourced from Shein and similar ultra-fast-fashion suppliers through TikTok Shop, sometimes spinning emotional narratives about handcrafting the items themselves to build parasocial trust with viewers.

The ethical problems stack up quickly. Real Black creators and small-scale artisans lose sales and attention to fraudulent competitors who have no actual stake in the communities they're performing. The use of AI-generated Black likenesses without consent echoes a long history of profiting from Black identity while excluding Black people from the gains. And for consumers, the entire interaction is a deception — they believe they're supporting a real person when no such person exists.

This isn't happening in a vacuum. TikTok has aggressively expanded its Shop and live commerce features, particularly in the U.S. market, modeling itself on the wildly successful livestream shopping ecosystems that dominate in China. The low barrier to entry for TikTok Shop sellers — combined with limited identity verification — has made it a fertile ground for exactly this kind of grift. Shein itself has faced repeated criticism over intellectual property theft, labor conditions, and environmental impact, and its role as the product supplier of choice for these AI-powered scam operations adds another layer to its ongoing controversies.

The tools enabling this behavior have become remarkably accessible. Platforms like Midjourney, Stable Diffusion, and AI avatar video generators such as HeyGen allow anyone to produce a convincing synthetic human presence at minimal cost. What once required significant technical skill can now be done in minutes. The result is a new category of fraud that sits at the intersection of generative AI misuse, identity theft, and platform commerce abuse.

Responses from TikTok and other platforms have so far been inadequate. Calls for mandatory AI-generated content disclosure, stronger seller verification, and clearer community standards around synthetic personas have grown louder, but enforceable regulation remains far behind the pace of the technology. The FTC and other consumer protection bodies have flagged AI-enabled deception as a priority, but the sheer volume of content makes enforcement deeply challenging.

What makes this case particularly pointed is who bears the cost. The creators being undercut are often independent artists and small business owners. The identities being appropriated belong to communities that have historically had less institutional recourse when exploited. As generative AI tools grow cheaper and more capable, the question of who gets protected — and who gets used — will only become more pressing. Platforms that profit from commerce activity have a direct financial incentive to act, but whether that incentive will translate into meaningful change remains, for now, an open question.