Today I want to share a somewhat scary trend that I noticed recently. Artificial intelligence is a cool thing, it helps me generate ideas for 3D printing and designs for decorative wall stickers. But now AI has started working against me too.
A new scam scheme has appeared on marketplaces, and I want to tell you about it so that you are armed if you also sell physical goods. Here is the point. I send a product to a client in perfect condition, a sticker or a 3D printed item. The buyer receives it, takes a photo, and then uploads this picture to a neural network, for example, ChatGPT or Nano Banana. And they write a prompt saying, Break this item, or, Add a stripe to the sticker.
The AI draws cracks on a mask, for example, makes chips, and adds stripes to the sticker to imitate a printing defect. The client sends me this photo demanding a full refund. And often with threats saying, Return the money immediately, or I will leave a negative review. Before, especially during the holiday rush, customer support often gave in. Seeing the photo, they got scared of a bad review and immediately issued a refund without even looking into it. But when such requests started coming 2 or 3 times a month, and they looked suspiciously identical, we went to read Reddit and Google together. It turned out to be a massive problem.
How we fight this now. First of all, we carefully examine the photo for artifacts. Distorted text. If there is writing on the background or the product, AI often turns it into blurred nonsense. Illogical damage. A plastic mask printed on a 3D printer cannot break like glass. If there are too many chips or they are unnatural, this is a red flag. Strange edges. The places where the item is broken often turn out too smooth or round with AI. Mismatched details. The broken pieces do not visually fit the place of the break.
Moreover, we even started using AI itself for protection. We upload the client's photo to Gemini or ChatGPT and ask, Is this a real photo or AI generation? And if the neural network confirms that it is a fake, we ask it to explain in detail and point by point why it made such a decision. Then we use this information and try to convey these arguments to the buyer first, again writing the response using AI. If this does not work, the tactic is as follows. We do not issue a refund immediately.
We write to the client, No problem, we are ready to give you a full refund. Send us this broken item back, we will inspect it and return the full amount or send a new one. We will also pay for the return shipping. An honest buyer will send it without questions. But a scammer usually disappears because they have the whole product in their hands and nothing to return.
The only problem that remains is negative reviews. A scammer can write something nasty out of revenge, and marketplaces like Etsy are very reluctant to delete them. In 99 percent of cases, they ask to resolve the issue with the buyer, and they do not hear our arguments about AI fakes yet. We just have to write a polite public reply for future clients, explaining that we offered a full refund provided the item is returned. That is the war with technology. If you have any ideas or experience on how to delete such reviews or fight these generated broken goods, please share, I will be very grateful. We need to stop this madness somehow.
