Scammers Use AI to Create Fake Photos and Request Refunds in China

Imagine buying live crabs online. They arrive fresh, healthy, ready for the pot. But instead of cooking them, you take a photo of the intact product, use an artificial intelligence app to generate an image of the dead crustacean, and send it to the seller demanding your money back. You end up with both the crabs and the refund.

This isn't a science fiction script. It's the newest frontier of digital fraud spreading across e-commerce platforms in China, an ecosystem where the "refund without return" policy for low-value items has opened a loophole exploited with impressive ingenuity.

The Fraud in a Few Clicks

The mechanism is frighteningly simple. After receiving a product in perfect condition—be it a bedsheet, an electronic device, or a piece of fruit—the scammer uses one of the many available generative AI applications. They upload a real photo of the item and type a simple command: "tear this fabric," "break this phone screen," "make this banana look rotten."

In seconds, the AI delivers a hyper-realistic image of the damaged product. This fabricated proof is then sent to the customer service of giants like Taobao or Pinduoduo. To the seller who receives the complaint, the image looks like irrefutable evidence that something went wrong with the delivery. The refund is processed, and the fraudster celebrates another successful scam.

What once required damaging the product itself or having advanced Photoshop skills is now within anyone's reach with a smartphone. It's the democratization of digital crime.

From Isolated Cases to a Parallel Industry

What could have been seen as an isolated trick has quickly turned into a trend. On Chinese social media, like Xiaohongshu (a local version of Instagram), tutorials teaching how to pull off the scam are multiplying. Users share tips on which AI apps work best and how to make the fake images even more convincing.

The problem has escalated to the point where online sellers are starting to publicly vent about their losses. Small shop owners, who operate on tight margins, are the most vulnerable. For them, a wave of fraudulent refunds can mean the ruin of their business. Trust, a fundamental pillar of e-commerce, is being systematically eroded by images that lie with perfection.

The Digital Arms Race

The platforms, of course, are not standing still. A true technological arms race has begun behind the scenes. If scammers use AI to create fraud, e-commerce companies are racing to develop their own AIs capable of detecting these fakes. Image forensic analysis tools look for digital artifacts, lighting inconsistencies, and other subtle signatures that generative models leave behind.

But it's a cat-and-mouse game. With every advance in detection, more sophisticated AI models emerge, capable of creating images that are increasingly indistinguishable from reality. It's a battle fought in pixels, where the line between true and false becomes blurrier every day.