Cover News Reporter Meng Mei Ouyang Hongyu

Short videos with beauty filters, over-beautified planting notes, and AI-synthesized fake product pictures have become one of the biggest problems in the e-commerce industry in recent years...... "wrong version".

針對行業亂象，3月27日，淘寶正式發佈新規，率先啟動AI假圖治理，嚴厲打擊利用AI假圖欺騙消費者、侵害原創品牌商家的違規行為，保護原創商家和消費者權益。同時，淘寶向全行業發出倡議：規範使用AI生成圖片，保障商品信息真實，一起加入對AI假圖的全面治理。

According to reports, the platform will clearly identify typical AI fake picture scenarios such as "inconsistent material or style of goods", "distortion or exaggerated effect", "distortion of goods caused by non-existence of scenes", "obvious distortion of cutout textures" and other AI fake images, and launch a recognition model, and comprehensively govern the platform's AI fake pictures through measures such as "source interception", "stock cleanup" and "identification reminder".

具體而言，在商品發佈端，實時檢測和攔截失真圖片。此外，在商品詳情頁還將上線AI合成提示標識，以保障消費者的知情權。截至目前，平臺已累計攔截過度PS、AI誇大的假圖近10萬張。

The so-called AI fake image refers to the use of AI and other technologies to synthesize and present significantly distorted and inconsistent product renderings, including but not limited to AI fictitious product effects, plagiarism and splicing of original styles, tampering with product materials or functions, and forging and splicing false scenes. The existence of a large number of AI fake images not only misleads consumers, infringes on consumers' right to know, but also disrupts the business ecology, causing bad money to drive out good money.

AI technology has advanced by leaps and bounds, and AI-generated product maps have been widely used by merchants. In response to this phenomenon, recently, the Cyberspace Administration of China and other four departments jointly issued the "Measures for the Identification of Artificial Intelligence-Generated Synthetic Content" to control AI fraud from the source. As AI applications enter the deep waters, there is an urgent need to standardize applications. "It's a good start to manage AI fake images. However, promoting AI for good requires the concerted efforts of the whole society, and sustainable development can only be achieved by restricting innovation on the track of authenticity and integrity. An AI researcher said.

"AI may not have a value judgment, but AI consumers must have a judgment and the platform must have a value orientation." In this regard, Taobao emphasized that the development and use of AI should improve consumer experience, promote originality, and protect the rights and interests of legitimate merchants, rather than becoming a tool for illegal profits in the black and gray industry. "Problems such as cross-platform picture theft, copycat copying, and excessive beautification are severe, and they cannot be cured by one platform alone. We have also issued an initiative to the whole industry to standardize the use of AI-generated images to ensure the authenticity of product information and return the industry to quality competition. ”