Taobao and Tmall have introduced a new AI-powered model designed to identify counterfeit or artificially altered images submitted during after-sales disputes. The system is now available to merchants with store ratings above 4.8 points, who can use a dedicated feedback channel to report suspicious product images that may have been digitally manipulated.
This initiative is part of a broader effort to combat the growing issue of buyers using AI-generated or edited images to fraudulently claim refunds. The platform has established an AI-based identification and governance system to handle disputes, refund requests, and appeals, aiming to create a more stable and fair e-commerce environment for sellers.
Merchants can report suspected fake images directly within the Qianniu Wangwang chat interface by right-clicking on an image, selecting "Report False After-Sales Evidence," and submitting the relevant order and scenario details. If the AI model confirms the image is fraudulent, it will directly influence the dispute outcome and alert the merchant.
The detection model utilizes Alibaba Group's latest AI-generated image identification technology, trained on a large dataset of authentic and AI-altered images across various industries. It can identify fully AI-generated images, images with visible watermarks or logos, and genuine images that have been artificially edited. The feedback feature is currently available to high-rated stores and will gradually expand to more merchants as the model's accuracy improves.
According to platform data, the merchant credibility system launched last year has helped sellers avoid over 4 billion yuan in losses. Stores with ratings above 4.8 points have seen transaction growth rates 2.2 times higher than those with ratings between 4.5 and 4.8 points. Earlier this year, Taobao and Tmall announced plans to enhance their business environment by 2026, including the rollout of the AI fake image detection model and a mutual trust mechanism for return discrepancies.
Comments