70 million fake products: AI fakes flood TikTok Shop

 

Fake product listings are not a new phenomenon. As soon as copy-paste was invented, fraudsters discovered their ability to infringe on any brand’s IP rights by simply copying its content and posing it as their own.

 

Unfortunately, fraudsters have gotten more sophisticated over the years and now, thanks to AI, can produce incredibly authentic-looking product listings at the touch of a keyboard.

 

 

AI-generated fake product listings on TikTok Shop

TikTok Shop and Amazon both report the appearance of very convincing brand content, including ads, reviews, product listings, and even influencer videos on their platforms. The purpose of this content is to convince shoppers to buy the products. However, here’s the catch: these products don’t exist, and once buyers have parted with their money, they’ll never see anything in return.

 

All of this content is generated by AI, allowing fraudsters to create high-quality, compelling, and varied content to deceive shoppers. Gone are the lame visuals and typo-ridden texts of old; with a few well-worded prompts, fraudsters can create flawless product pictures, convincing marketing texts, authentic-looking reviews, and much more.

 

In many cases, even the brands themselves are fake, which means that no direct case of copyright infringement arises (unless the AI relies too much on an existing brand’s IP-protected assets).

 

The various types of fake content can even be combined freely, creating the image of a well-established brand with a legitimate, exciting product.

 

"It's organized crime, to be honest," said Nicolas Waldmann, the spokesperson in charge of TikTok Shop’s governance and external affairs team. "They're trying to basically go through and sell, and of course, never deliver anything, and then run with the money."

 

In response to this and other threats, TikTok Shop has removed 70 million product listings and banned 700,000 sellers in the first half of 2025.

 

Cybersecurity firm Trend Micro researched scams based on AI-generated fake content and found that 16.43% of all these scams involve fake merchandise. The overwhelming majority, 77.27% fall into the category of romance impostor scams.

 

Discover how AI will revolutionize the entire e-Commerce industry

 

Find out how AI can be enlisted to strengthen the protection of your IP rights

 

 

TikTok Shop’s rapid expansion and resulting vulnerabilities

The e-Commerce entry of popular video-sharing social media platform TikTok caused serious waves in the world of online shopping.

 

In the few years since its US launch in 2023, TikTok Shop is now generating a monthly $1.1 billion GMV in the country. 71.4 million US-based consumers buy products on TikTok Shop, which corresponds to 45.5% of American TikTok users.

 

Globally, over 15 million sellers are active on the platform, generating around $33.2 billion GMV all over the world.

 

However, such rapid expansion, combined with TikTok Shop’s open marketplace model, resulted in serious vulnerabilities, especially when it comes to fraudulent content appearing on the platform. This is further aggravated by TikTok’s very purpose - creating and sharing short, attention-grabbing videos online.

 

Discover how TikTok Shop enters EU e-Commerce in Spain, France, Germany, Ireland, and Italy

 

Given TikTok Shop’s creator-driven commerce strategy, fraudsters have an easy time hiding their AI-generated scams amidst the myriad of genuine videos. There’s no shortage of issues: AI deepfakes of real doctors selling bogus health supplements, faked influencer videos advertising counterfeit products, and AI-generated campaigns promoting cryptocurrency scams are just a few of the most recent scandals involving AI-based fraud on TikTok.

 

 

Illustration of an AI-generated influencer promoting a product

 

 

AI has reached a level where moderation teams on popular platforms like TikTok and Amazon have a hard time distinguishing genuine content from AI-generated fakes, allowing an increasing number of fakes to appear and wreak havoc on platforms.

 

 

How TikTok responds to AI-related threats

According to Nicolas Waldmann, TikTok Shop uses a combination of human and AI-assisted moderation to filter out fraudulent listings and accounts. The company has in-house tools for this purpose, but it also works with external partners to achieve the best results.

 

These external partnerships assist TikTok Shop in various tasks, including the authentication of pre-owned luxury goods.

 

Find out how eBay authenticates designer sneakers

 

Discover Amazon’s venture into the world of second-hand luxury goods

 

As online brand protection experts, we believe that one of the best ways to exclude fraudulent product listings is a stricter vetting process for sellers. e-Commerce platforms should introduce a more thorough verification system to identify sellers before they’re allowed to sell, ensuring that applicants are genuine merchants and not counterfeiters.

 

Brand partnerships are another great way to establish that only authorized sellers of a certain brand can access the platform, eliminating bad quality fakes and lookalikes.

 

As a general principle, online marketplaces should always devise and enforce a strict IP protection policy to curb the amount of infringing content appearing on their platform.

 

 

Amazon’s measures against AI scams and fake products

Amazon is also affected by the sudden rise of AI-generated fake content on its platforms. The e-Commerce giant, similarly to TikTok Shop, decided to fight fire with fire and lined up a collection of AI tools to detect and combat the threat of infringing content.

 

Although none of the reports elaborate too much on the nature of these tools, they seem to be able to identify AI-powered manipulation in images, detect fake reviews, and distinguish between genuine and fake storefronts.

 

As the measures were set up in early 2025, there are already some results to report. Amazon claims that its AI tools have found and deleted 99% of suspicious listings before they were detected by brands. This led to a 35% reduction in the number of infringement claims raised by brands compared to similar data in 2020, even though the amount of infringing listings is on the rise.

 

Another positive note is the fact that Amazon’s initiative to completely eliminate counterfeits available on the platform, Project Zero, has reached an important milestone. The program, introduced in 2019, has now helped over 35,000 brands to proactively block fake product listings. The program uses automated protection as well as instant removal tools that enable brands to quickly and efficiently demand the removal of fake product listings from the platform.

 

 

Shared challenges: scalability, accuracy, and global enforcement

While fake content is nowhere near new, the recent flood of AI-generated counterfeit listings seems to have hit e-Commerce platforms harder. This may very well be because of the scale and sophistication of the attacks.

 

There’s also another factor: since genuine brands may also use AI-generated content in their legitimate marketing material, platforms can’t simply discard anything that screams AI. This further complicates the issue of telling genuine from fraudulent content.

 

To improve the accuracy of their response, platforms should consider using a hybrid approach where human teams work alongside AI-powered tools. Such a balance between automation and human expertise allows the strengths of each resource to compensate for the weaknesses of the other.

 

Another serious issue is the fragmented response to a global threat. Fraudsters often belong to networks that launch complex, coordinated attacks on several online marketplaces and social media platforms at once. However, platforms rarely coordinate their response, which significantly hinders their efficiency.

 

Therefore, global cooperation and enforcement are needed to effectively address the threat of both AI-powered and traditional scams.

 

Find out how a clean marketplace contributes to your brand’s success

 

Marketplace Monitoring 2.0: How data analysis and automation affect brand safety

 

Learn all about these Chinese social media platforms’ e-Commerce ventures

 

 

The role of proactive brand protection services

Whether generated by AI or traditional means, online monitoring reliably detects infringing content. This is why it’s crucial to set up a comprehensive online brand protection program and use it proactively to nip threats in the bud before they have the chance to escalate.

 

Continuous cross-platform monitoring is a must. One-off searches (e.g., via Google Lens) simply don’t cut it anymore in our current era of rapidly evolving threats. Instead, you need to set up regular monitoring intervals across online marketplaces, social media platforms, domains, and imagery. This is the only way to detect connections between seemingly independent threats (for example, fake single webshops advertised on social media) and deal with them effectively.

 

To best counter the threat of AI-generated fakes, modern brand protection services have to include AI tools as well. However, this doesn’t mean that you can solely rely on AI. On the contrary, as outlined above, human expertise and AI-powered efficiency need to complement each other to eliminate each other’s weaknesses.

 

Discover how we incorporate AI features into our beloved image search tool, infrimage

 

Once infringing content is detected, it’s imperative to have it removed from the internet as soon as possible. This is why a flexible and versatile enforcement service needs to be a part of your brand protection program. With infringing content eliminated from TikTok, Amazon & Co., your genuine content can finally shine in the clean marketplace it deserves.

 

 

Key takeaways for your brand

As AI-driven e-Commerce frauds transform the landscape of threats, constant adaptation and continuous monitoring are required to keep your IP rights safe from infringement.

 

Luckily, brands don’t face this challenge alone; online brand protection experts like globaleyez are readily available to assess the threats facing your brand and create the most effective personalized response. As you can see, even large marketplaces like Amazon rely on the expertise of external partners to effectively counter the arising threats of AI-powered infringements.

 

Such a strategic partnership with online brand protection providers ensures a scalable, data-driven, and highly effective response perfectly tailored to your brand’s needs.

 

Contact us if you’d like to know more about the threat of AI-generated infringements and how best to protect your brand against them!