How AI is used to generate fake product listings and reviews on marketplaces

Artificial intelligence has long been a business tool: it is used to write product descriptions, process images, and respond to customers. However, unscrupulous sellers are also increasingly using these same technologies—generative AI makes it possible to prepare a product listing in just a few minutes that is almost indistinguishable from the original and generate dozens of plausible reviews. Let’s examine how this works and why brands should rethink their approach to intellectual property protection.

Counterfeits have become harder to recognize

Previously, violations could be identified by exact matches of names or photos, as counterfeit listings gave themselves away through low-quality images, description errors, and sloppy presentation.

Today, using generative AI, sellers create professional content in minutes that is nearly indistinguishable from the original. AI rewrites product descriptions and modifies images based on original materials so that the listings do not appear to be direct copies of one another. Meanwhile, the trademark and brand identity elements are altered only slightly, remaining recognizable to consumers; consequently, a simple search for duplicates no longer detects such listings. Now, identifying violations requires analyzing a multitude of more subtle indicators.

How it looks in practice

Let's take a typical scenario that brand protection teams face today. An unscrupulous seller takes photos and descriptions of the original product and runs them through a generative model: the text is rewritten to be close to the original but with different wording and structure; in the photos, the angle, background, and color scheme are slightly changed, and the logo on the packaging is blurred or cropped. Formally, the listing is not an exact copy—neither the text nor the image matches the original byte-for-byte, and while the buyer is essentially looking at a counterfeit, they involuntarily associate it with a well-known brand.

Next, this listing is multiplied: several different versions of descriptions and photos are generated, published under different sellers, and then "boosted" with reviews—dozens of short but differently phrased comments about product quality, fast delivery, and accuracy of the description. No automated marketplace filter searching for duplicate text or matching images flags such a listing: each copy is unique at the level of phrasing and pixels, even though the entire group of listings clearly originates from a single source.

To find such a group of listings manually, a brand protection specialist must look not for an exact match, but for indirect signs: similar frame composition, turns of phrase characteristic of a specific seller, and repeating patterns in reviews. In practice, this means that a single case takes many times longer than it did a few years ago, and the number of such cases is increasing.

By what signs can a group of AI cards be identified?

Each card in such a group is unique at the level of text and pixels, but they retain a common origin, which manifests in indirect indicators:

- Similar composition. The camera angle, product placement in the frame, light balance, and color scheme are repeated, even if the background and small details are changed.

- Identical AI "errors." Similar distortions of the logo and inscriptions on the packaging, and recurring image artifacts characteristic of generative models.

- Unified description structure. The texts are written in different words but organized in the same order: the same highlights, the same blocks, and similar paragraph lengths.

- Connections between sellers. Formally different stores match in terms of registration dates, assortment, prices, delivery terms, and other profile data.

- Synchronized publication. The cards appear almost simultaneously, and each quickly gains a rating.

- Similar reviews. Short comments about quality, delivery, and description accuracy differ in wording but are similar in structure and the range of topics.

None of these indicators individually proves a violation. Their combination becomes significant, and finding it requires comparing dozens or hundreds of cards simultaneously, which is almost impossible to do manually.

Reviews you can't trust

A separate issue is the automatic generation of reviews. Modern language models create large volumes of natural-sounding comments, varying in style and length, which are difficult to distinguish from real ones—and this is not just an assumption, but a result confirmed experimentally.

A 2025 study showed that people are no longer able to distinguish real reviews from AI-generated ones—the average recognition accuracy was only 50.8%, which is practically equivalent to random guessing. Moreover, language models themselves also fail at this task and show results no better, and sometimes worse, than humans. In other words, neither humans nor algorithms today can reliably separate a real review from a generated one, meaning that automated marketplace filters relying solely on text analysis catch such fakes only slightly more effectively than random selection (see Meng W., Harvey J., Goulding J., Carter C. J., Lukinova E., Smith A., Frobisher P., Forrest M., Nica-Avram G. "Large Language Models as 'Hidden Persuaders': Fake Product Reviews are Indistinguishable to Humans and Machines").

Such reviews are used to artificially increase trust in a product and promote listings in marketplace search results. As a result, a high product rating is no longer a guarantee of quality, and negative reviews can be a tool for unfair competition. For the buyer, this means that familiar indicators for choosing a product are ceasing to work.

What this means for rightsholders

The use of AI makes counterfeiting not only more widespread but also more convincing, which increases the risks for brands:

• the number of violations is increasing, and they are becoming harder to track manually; • fake product listings are harder to distinguish from original ones, even upon close inspection; • inflated ratings and fake reviews nudge customers into buying counterfeits, leading to disappointment with low-quality goods that are then associated with the brand itself rather than the infringing seller.

Ultimately, both sales and reputation suffer, and the damage to reputation may not become apparent until much later.

What a brand can do right now

Automated monitoring addresses the challenge of scale, but the effectiveness of brand protection also depends on how well a brand prepares in advance. There are several steps that require little investment yet can significantly simplify the fight against counterfeit product listings.

1. Check the scope of your trademark registration. Make sure your trademark is registered in the relevant classes of the Nice Classification (NCL) for your business activities and in the countries where your products are sold. Without registration, filing complaints with marketplaces or taking legal action can be more difficult. AI-generated copies often use slightly modified versions of a trademark, so it is also worth registering its key variations.

2. Record your trademark in the customs register. If counterfeit goods are imported from abroad, recording your intellectual property in the Russian Customs Register of Intellectual Property Objects (TROIS) can help stop shipments at the border before counterfeit listings appear on online marketplaces.

3. Compile a list of authorized sellers. Clearly identify who is authorized to sell your brand's products and publish the list on your website. This makes it easier for both customers and marketplaces to distinguish genuine products from counterfeits and helps your team identify infringers.

4. Keep reference materials. Store original photographs, product descriptions, packaging designs, and other brand assets in one place, along with their creation dates. This helps demonstrate that a copy was based on your materials, even if it does not reproduce them verbatim.

5. Document infringements and preserve evidence. Save screenshots showing the date and page URL. For significant cases, consider commissioning a notarial examination of the webpage or using specialized digital evidence-capture services. It is best to collect evidence before a listing is removed or modified.

6. Monitor reviews as well as product listings. A sudden surge in similar positive reviews on a newly created listing or a spike in negative reviews on a genuine product may indicate coordinated review manipulation. Flag such cases and take them into account when filing complaints.

7. Use the platforms' own protection tools. Many marketplaces offer brand protection programs and dedicated complaint submission forms for intellectual property rights holders. Registering for these programs can reduce the time it takes to process complaints. These measures establish a solid foundation, but they do not eliminate the main challenge: manually checking hundreds of product listings and tracking changes to them is simply not feasible.

Why system monitoring is needed

Periodic manual checking of marketplaces can no longer keep up with the scale of the problem: employees would have to analyze hundreds of listings, compare images, and track changes almost in real time. Even a strong team will inevitably miss some violations.

Therefore, companies are switching to continuous automated monitoring with prompt documentation of violations, filing of complaints, and analysis of repeat cases of illegal trademark use. Solutions like ZIPDetect help not only to find suspicious listings but also to build the entire brand protection process—from monitoring to legal support.

The earlier a company implements such tools, the higher the chance of noticing a violation in time and not losing customer trust.