“以前模特图虽然会狠狠P图，但至少能看面料和质感，现在直接用AI，请问是要卖给AI吗？”不少网友在社交平台吐槽，电商AI模特图“失真”“无法判断版型”。双十一临近，电商AI模特展示乱象引发热议。澎湃新闻调查发现，有的店铺AI模特手部扭曲、长出三条腿，评论区甚至出现“三只脚”的AI好评晒图。破绽如此明显，AI生成标识却近乎隐形，真假难辨，消费者看得一头雾水。
一家网店中的女童模特图，手部姿势明显违背人体结构。
乱象背后，是一条成熟的AI模特产业链。实拍一件服饰，模特、场地、拍摄打包下来至少两三百元；AI生成模特图最低仅6元一张，成本优势一目了然。在网上，AI女装生图提示词合集、带货工作流教程低价售卖，动辄上万份销量。商家发来一张衣图，半小时就能拿到AI成片，实拍却往往要等一两天。一边是商家降本增效的现实诉求，一边是消费者“货不对图”的持续吐槽，部分平台相继出台规则整治AI假图、明确AI标识要求。
首先应该承认，商家使用AI模特图本身并不违法。用AI生成虚拟模特用于商品展示，属于人工智能正常应用，现行法律并未禁止。技术本身没有原罪，即便是真人实拍图，也可能过度修图、脱离实物；AI图只要如实还原商品，同样具备正当性。诚如不少网友所言：“真人展示也未必完全真实，我不排斥做得好的AI。”
AI不是原罪，隐瞒却是陷阱。法律约束的从来不是技术工具，而是误导与欺诈。商家使用AI模特是否构成虚假宣传、消费欺诈，评判标尺从来不是“有没有用AI”，而是展示画面和实物是否一致、有没有刻意隐瞒信息。只要客观呈现面料、版型，不虚构品质、不隐瞒瑕疵，主动标注AI属性，不做虚假诱导，就不触碰消费欺诈红线。
反过来，倘若AI效果图和实物差距巨大，突破合理美化的边界，再叠加客服谎话搪塞、刻意抹除AI标识等行为，就极易被认定为消费欺诈。按照《消费者权益保护法》第五十五条，商家要承担“退一赔三”的惩罚性赔偿。AI从来不是商家的免责金牌，更不是欺诈的隐身衣。谁试图拿AI当挡箭牌，谁就要为虚假宣传买单。
《人工智能生成合成内容标识办法》早已明确，AI生成图文视频，应当设置清晰可感知的标识；隐而不标、足以误导消费者，便是侵害知情权。平台同样负有核验义务，对检出合成痕迹却未声明的内容，应当及时标注提示，当好消费信任的“守门人”。
AI模特“攻占”双十一，大可不必因噎废食。工具本无善恶，责任在于使用工具的人。规范用好AI，守住不欺诈消费者这条底线，AI模特才能走出真假难辨的舆论漩涡，褪去“造假工具”的负面标签，真正扎根实体经济，服务消费者，成为电商行业可信、好用的数字帮手。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
"Although the models used to be tough, they could at least look at the fabric and taste, and now use AI directly. Would you like to sell it to AI?" There are a lot of netizens throwing up on social platforms, and the electrician AI model is "unable to judge the version." Two elevens approaching, the electrician Ai model showing a hysteria sparks debate. A press survey found that some shop models had twisted hands and raised three legs, and even “three feet” in the comment area had a good view of AI. The cracks are so obvious, and the AI generation label is almost invisible, and it's so hard to tell that consumers see a fog.
The modeling of girls in an Internet store, with hand positions clearly contradicting the human structure.
Behind the mess is a mature AI modeling business chain. It's at least $2,300 for a dress, a model, a venue, and a camera, and it's only $6 for an AI model, with a clear cost advantage. On the Internet, the AI Women's Fashion prompt collection, with the goods workflow, is sold at a low price, with the highest sales volume. The merchants send a picture of their clothes, they can get an AI film in half an hour, but they often wait for a day or two. On the one hand, there is a real demand from businesses to reduce efficiency gains, and on the other hand, there is a constant blundering of consumer “no-goods” maps, with some platforms successively putting in place rules to correct the AI false maps and clarify the AI marking requirements.
First of all, it should be recognized that the use of the AI. Model is not illegal. The use of AI to generate virtual models for the display of goods is a normal use of artificial intelligence, which is not prohibited by current law. Technology itself is not an original crime, even if it is a real one, it can be overtly modified and dematerialized; AI maps are equally valid as long as they are actually returned to the original commodity. As many netizens have said, “The real person's presentation is not necessarily entirely true, and I don't exclude the good AI.”
AI is not a crime, but a trap. The law was never a technical tool, but a misleading and fraudulent one. The use of the AI model by a commercial would constitute false propaganda, consumer fraud, and the ruler would never be “use or not”, but would show whether the image and the object were consistent and whether the information was deliberately concealed. The red line of consumer fraud is not touched as long as the fabric, layout is presented objectively, the quality is not made up, the defects are not concealed, the AI attributes are indicated on their own initiative and no false inducement is made.
Conversely, if there is a wide gap between AI impact maps and physical objects, breaking the reasonably beautified border, and adding false statements, deliberately removing the AI logo, etc., it is very easy to be considered consumer fraud. Under article 55 of the Consumer Protection Act, a merchant is liable for punitive damages of “one to three”. AI was never a merchant's immunity gold medal, much less a fraudulent cloak. Whoever tries to use AI as a shield will have to pay for false propaganda.
The Artificial Intelligence Synthetic Content Marking Scheme is well established, and AI generates graphic video, which should be clearly and sensitively marked; it is a violation of the right to know when it is invisible and misleading to the consumer. The platform is equally under an obligation to verify that undeclared traces of synthetics are detected, and should be prompt to be a “gatekeeper” of good consumer confidence.
The Ai model “takes” two elevens, so there's no need to choke. Tools are not good or bad, and the responsibility lies with those who use them. By regulating AI, and keeping the bottom line of consumer fraud, the AI model will be able to emerge from the real and unconvinced vortex of public opinion, remove the negative labels of “false tools”, truly root the real economy, service consumers, and become a credible and useful digital helper in the electric industry.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.