“Editing my photos to get my money back on DoorDash”… (Left: real photo. Right: AI-edited photo)
AI technology is being used to manipulate product photos, with customers claiming orders arrived damaged and fraudulently demanding refunds.
Chinese e-commerce platforms are seeing a rise in fraudulent refund requests, driven by artificial intelligence (AI). Consumers are sending AI-generated images of damaged products to get the money back.
The problem intensified during this year’s Double 11 shopping festival in November, when several online merchants reported having received manipulated photographs which supposedly showed moldy fruit, cracked mugs or torn clothes.
Some buyers claim to take real photos of the items they receive and use AI tools to digitally alter them. A fruit seller said customers used AI to turn fresh fruit into rotten fruit, while an electric toothbrush seller received an image of a toothbrush edited to look rusty. “They come to me with problems that go against common sense,” the shopkeeper told .
A clothing retailer reported a case in which a customer requested a full refund for a dress, claiming that the collar was frayed. The attached photo, however, showed telltale signs of AI alteration, such as inconsistent lighting and artificial margins, which led the team to reject the request.
In another incident, a mug seller noticed that, despite secure packaging, a buyer sent a photo of a mug with cracks that looked like cobwebs. When asked to present video evidencethe buyer quickly withdrew the complaint. An AI detector later suggested a 92% probability that the image was artificial.
Sellers claim that existing rules often favor consumers, making it easier for people with bad intentions to abuse the system. In response to growing concerns, major Chinese platforms including Taobao and Tmall eliminated the “refund only” option earlier this year and introduced a buyer credit rating system based on purchase history, refund patterns, and seller feedback.
The phenomenon is not limited to China, with several North Americans use the same method with food delivered at home through platforms like DoorDash and some even justifying the practice because they ran out of vouchers to buy food during the Government shutdown.
Editing my pictures so I can get my money back on DoorDash
— ☢ (@King_Sukunaaa)
But companies aren’t the only ones misusing AI: some marketers also use generative tools to beautify product adscreate virtual models or fabricate backgrounds, according to industry experts. There are already several reports of this type of situation in .
China also implemented new AI identification rules on September 1, requiring the Explicit and implicit labeling of AI-generated content. Some platforms have implemented AI detectors to flag unlabeled images, although experts say accuracy is still a challenge due to the rapid evolution of generative technologies.
