Recently, some e-commerce platforms have announced the launch of AI fake image governance on all platforms, bluntly saying that AI fraud has become an accomplice to inferior merchants and goods, harming intellectual property rights and infringing on the interests of consumers.

Jiefang Daily and Shangguan News reporters investigated and found that not only should the "seller show" not be trusted, but AI fraud also appeared in the "buyer show"; Not only the pictures of the "buyer show" and "seller show" may be beautified after being blessed by AI, but the "seller show" and "buyer show" displayed in live broadcasts and videos also have moisture.

On many e-commerce platforms, AI fraud can be seen everywhere. In the product introduction process, some merchants use AI technology to generate fake pictures in order to attract consumers' attention. Xu is not divided into counterfeit merchants "not skilled", a little careful inspection, you can find the clues-

Different merchants use the same "AI model" for "seller show", and the posture and expression of the model are exactly the same, and even because the retouching is not careful, some models have missing arms and legs;

The expressions and postures of the models in the AI fake "seller show" (most likely generated by AI) are exactly the same.

The model with the same expression and posture has missing arms and legs. (Source: AI fraud picture disclosed by the e-commerce platform)

In the background of the AI-generated "seller show", there are many details that defy common sense, including unrecognizable fonts, seasonal features that contradict the model's dress, and so on.

There will be a lot of details in the AI-generated "seller show" that defy common sense. (Source: AI fraud picture disclosed by the e-commerce platform)

With the upgrading of AI technology, it is not uncommon for short videos and live broadcast room explanations to be faked. Some recorded "video explanations" use technically adjusted videos to deceive consumers; The live broadcast seller uses the real-time beautification function of the shooting software to sell the goods that have been modified by the software in the live broadcast room.

Similar flaws can also be found in AI fake videos. For example, when a seller showed a pair of belly pants that claimed that consumers could immediately become "small waist essence" after wearing them, there were multiple stutters on the video screen. After identification, it turned out that the picture was incoherent due to editing, and the "small waist essence" was a picture after technical processing.

What is even more alarming is that the "buyer show" that could have provided consumers with reference has also shown traces of AI fraud. In the comment area of some products, there are many "buyer shows" with similar content; Or different merchants sell the same product, and as a result, there is also a "buyer show" with the same picture in their respective evaluation areas.

For example, in the review section of a clothing store, there are many "buyers" who show photos of themselves wearing certain clothing. Although the poster has hidden the face and other highly distinguishable details, after comparison, the shooting posture, product display method and even the shooting background are very similar, and it cannot be ruled out that it is the result of the seller's swiping comments after the AI batch generation of the "buyer show".

Some of the "buyer shows" seem to be posted by different users, but the details are strikingly similar, and some even directly intercept the "seller show". (Source: E-commerce platform evaluation area)

For example, in the evaluation section of a liquid foundation, there are multiple photos showing that the "consumer" has perfect skin condition and long-lasting makeup after using the product. But if you search the comments section of other foundations, you'll find plenty of the same "buyer's show" photos.

In the face of these "seller shows" or "buyer shows" full of counterfeit traces, real consumers are likely to encounter "mismatch".

All of the above has undoubtedly sounded the alarm for consumers who like to go online and go to the live broadcast room to "buy, buy, buy": "seller show" should not be trusted, and "buyer show" should not be fully trusted.

What should I do if the e-commerce ecosystem encounters the challenge of AI fraud?

China's Law on the Protection of Consumer Rights and Interests clearly stipulates that consumers have the right to know the truth about the goods they buy or use or the services they receive, and business operators shall truthfully and comprehensively display information about goods to consumers, and shall not make false or misleading publicity.

There is no doubt that the emergence of AI fraud has seriously violated consumers' right to know and the right to fair trade. When consumers make a purchase decision based on fake images or fictitious videos, they often receive items that are far from what they expected, wasting money and impacting the shopping experience.

More importantly, AI fraud appears in both the "Seller Show" and the "Buyer Show", indicating that the e-commerce ecosystem is encountering new challenges - AI fake images disrupt the normal market order, and those merchants who operate in good faith and use real pictures to display their products, as well as consumers who publish real consumer reviews, may be at a disadvantage due to the unfair competition of AI fake image merchants.

Taking the field of clothing e-commerce as an example, some original design brands take real clothing display pictures with their hearts, truthfully showing the material, version and details of the clothing. However, due to the imperfection of the "seller show", the sales volume may not be as good as those merchants who use AI fake pictures to exaggerate their publicity.

In the consumer evaluation process of the e-commerce ecosystem, the voices of consumers who truthfully evaluate the quality of goods are likely to be drowned out by the positive "buyer show" generated by AI batches, leaving other consumers at a loss.

The phenomenon of "seller show" and "buyer show" being fake at the same time may eventually lead to a vicious circle of bad money driving out good money and destroying the entire e-commerce ecological environment.

In the face of tens of millions of e-commerce merchants and a large number of "seller shows" and "buyer shows", it is obviously impossible to rely on manual identification of AI fraud. Defeating "magic" with "magic" has become a must-choose.

On the one hand, technology should be used to help consumers identify those "seller shows" involved in counterfeiting.

For example, according to typical AI fake image scenarios such as "inconsistent material or style of goods", "distortion or exaggerated effect", "distortion of goods due to non-existence of scenes", "obvious distortion of cutout textures", etc., the recognition algorithm can be set up to identify a large model for the platform online, and improve the efficiency of AI counterfeiting through measures such as "source interception", "stock cleanup" and "logo reminder".

Another example is that blockchain technology can be introduced to store the copyright of commodity pictures. When the merchant uploads the original shooting material, the platform implants an AI digital watermark, which can embed invisible logos without damaging the visual effect to avoid being stolen and tampered with by other merchants.

On the other hand, we should not let go of the "buyer show" fraud, but severely crack down on the phenomenon of swiping orders, brushing praise, AI buyer shows, and mismatching. For e-commerce platforms, the management of AI fake pictures and fake videos is only part of the anti-counterfeiting work, and what is more needed is to fight counterfeiting in an all-round way - not only pay attention to the authenticity of product introductions and user evaluations, but also strictly control the quality of goods and merchant marketing on the platform and in the live broadcast room, and stop fake promotions, swiping orders and other behaviors. At the same time, it is necessary to give more weight to the real "buyer show" to help other consumers understand the product more comprehensively and make the right decisions.

Source: Shanghai Net