According to Business Insider, TikTok Shop executive Nicolas Waldmann revealed the platform is fighting what he calls “organized crime” from AI scammers creating fake brands and products that never get delivered. The company uses a mix of AI and human moderation to combat these sophisticated fraud attempts, with Waldmann stating “We use AI to basically deal with AI.” In the first six months of 2025, TikTok rejected 70 million products and removed 700,000 sellers for policy violations. The platform drove $100 million in single-day sales on Black Friday last year in the US, making marketplace quality a critical priority. TikTok Shop has faced ongoing challenges since testing began in 2022, including sellers bypassing filters to offer prohibited items like THC syrup and sex toys.
Sponsored content — provided for informational and promotional purposes.
The AI Fraud Acceleration
Here’s the thing – this isn’t new fraud, but generative AI has supercharged it. Bad actors can now create convincing fake brands, product images, and entire storefronts in minutes rather than days. And they’re getting smarter about evading detection systems. TikTok’s approach of using AI against AI makes sense, but it’s becoming an arms race where both sides keep leveling up their technology. The real question is: can moderation teams ever truly stay ahead when the tools keep getting better and more accessible?
This Is a Much Bigger Market Problem
Look, this isn’t just a TikTok issue – every major e-commerce platform is dealing with this right now. Amazon, Shopify, even Facebook Marketplace are all seeing sophisticated AI-generated fraud. But TikTok Shop is particularly vulnerable because it’s built on that viral, impulse-buy mentality. When users are scrolling through endless videos and see something shiny, they’re more likely to click buy without doing due diligence. And these scammers know exactly how to exploit that psychological trigger. The $100 million Black Friday number shows how much is at stake here – if trust erodes, the entire shopping ecosystem collapses.
The Human-AI Moderation Dilemma
So TikTok’s using both AI detection and human moderators. That’s the right approach, but it’s incredibly expensive and difficult to scale. Rejecting 70 million products in six months is staggering – that’s basically an entire Amazon’s worth of fraudulent listings. And removing 700,000 sellers? That shows how widespread the problem has become. The challenge is that while AI can catch the obvious patterns, humans are still better at spotting the subtle tells of sophisticated fraud. But there’s never enough human bandwidth to review everything manually. It’s a constant balancing act between automation and human judgment.
What Comes Next in This Arms Race
Basically, we’re heading toward even more advanced detection systems. Think blockchain verification for luxury goods, better image analysis that can spot AI-generated product photos, and behavioral analysis that flags suspicious seller patterns before they even list products. But the scammers will adapt too. They always do. The real test for TikTok Shop will be whether they can maintain consumer trust while hitting those aggressive growth targets. Because if shoppers start doubting whether they’ll actually receive what they ordered, the entire TikTok commerce experiment could unravel pretty quickly.
