镜相工作室 . 以下文章来源于镜相工作室 ，作者镜相作者 商业世界的风向与人 违法与否，只在一句提示词之间。 文 ｜ 黄依婷 编辑 ｜ 卢枕 来源｜ 镜相工作室（ID：shangyejingxiang） 封面来源 ｜ unsplash 在很多人的印象里，造假是少数人的事。它需要专业的技术和硬件设备，需要行业人脉积累，造假者往往隐匿在灰黑产缝隙，是神秘的、恐怖的、没有道德 底线的，与大众保持着既远又近的距离。 但今年开始，一群人感受到了明显的变化。 11月，毛绒玩偶商家于瑾第一次遭遇AI假图仅退款。到货一周后，买家发来一张图申请仅退款，在店铺客服以人为损坏为由驳回后，买家申请平台介入并成 功拿回了50元。 但那张申请仅退款的图有一个不合常理的地方——玩偶柔软的裙边上，有类似陶瓷制品的坚硬裂痕。于瑾认定这是AI假图，发到社交平台吐槽。后来，在媒 体的介入下，平台自掏腰包，把这50元返还给了于瑾，用AI假图申请仅退款的买家却销声匿迹。 于瑾意识到，AI的普及让"羊毛党"的恶意退款成本更低了。一台手机、一句提示词就能生成一张以假乱真的图片，薅走一个价值一百多元的玩偶，放在二手 平台倒卖。"不用付出什么，也不会得到什么 ...

11月，毛绒玩偶商家于瑾第一次遭遇AI假图仅退款。到货一周后，买家发来一张图申请仅退款，在店铺客服以人为损坏为由驳回后，买家申请平台介入并 成功拿回了50元。 但那张申请仅退款的图有一个不合常理的地方——玩偶柔软的裙边上，有类似陶瓷制品的坚硬裂痕。于瑾认定这是AI假图，发到社交平台吐槽。后来，在 媒体的介入下，平台自掏腰包，把这50元返还给了于瑾，用AI假图申请仅退款的买家却销声匿迹。 文 | 镜相工作室 黄依婷 编辑 | 卢枕 在很多人的印象里，造假是少数人的事。它需要专业的技术和硬件设备，需要行业人脉积累，造假者往往隐匿在灰黑产缝隙，是神秘的、恐怖的、没有道德 底线的，与大众保持着既远又近的距离。 但今年开始，一群人感受到了明显的变化。 于瑾意识到，AI的普及让"羊毛党"的恶意退款成本更低了。一台手机、一句提示词就能生成一张以假乱真的图片，薅走一个价值一百多元的玩偶，放在二手 平台倒卖。"不用付出什么，也不会得到什么惩罚。" 今年更早时候，公关齐云也经历过类似的冲击。他服务的一家上市公司被AI生成的黑稿攻击，文中细节让他直言"离谱得很"。但受制于内容平台规则，他和 同事们不得不花整整两天时间，对文中细节一 ...

在一个"有图有真相""有视频有真相"的时代，当AI越来越聪明，你还能分得出真假吗？近来AI成为已经 成为"羊毛党"新工具——随着图片造假产业化，AI诈骗开始入侵电商领域。 技术热潮之下，生成式AI正在挑战并重建新秩序，但其被滥用的风险也在随之提升。从深度伪造引发 的内容失真，再到AI情感陪伴所触发的伦理困境，一系列典型案例不断进入公众视野，揭示了技术滥 用的现实隐患。 2025年12月18日下午，由南方都市报社、南都数字经济治理研究中心主办的第八届啄木鸟数据治理论坛 在北京举行。会上，南都发布了一份四万字的重磅报告《生成式AI落地应用：透明度测评与案例分析 报告（2025）》。 报告回顾了近些年118起生成式AI风险案例，聚焦生成式AI带来的社会信任挑战以及人机交互伦理争议 和治理痛点，并尝试提出切实可行的政策建议。 AI造假冲击个人权利、行业秩序和社会信任机制 司法场景中，海内外均出现律师利用AI生成虚假判例的情形。2025年10月北京通州法院披露，一名律 师在代理意见中引用了不存在的"（2022）沪01民终12345号"判决。法官因案号规律异常而识破，确认 该判例系AI凭空生成。该律师因未尽核实义务，被审 ...

据央视新闻报道，眼下，AI技术正赋能千行百业，为人们的工作、学习、生活带来很多便利。但最近，一些人竟然用AI制作虚假图片和视频向网店商家 索赔。来看江苏苏州一位大闸蟹经销商的遭遇。 买家发来"死螃蟹"图片视频 疑点重重 11月17日，在江苏苏州相城区经营大闸蟹生意的高女士，根据客户订单，发出了一箱8只装大闸蟹。第二天，广州的买家就收到了货，完成签收。就在签 收后10多分钟，高女士收到平台发来的客户反馈。 江苏苏州相城区大闸蟹经销商高晶：11月18日早上11点40分左右就签收了，11点53分就来反馈说8只螃蟹死了6只。 高女士对此也非常疑惑，立即联系消费者，请他们提供相应的证据。 很快，买家就发来了一张照片，画面显示，6只螃蟹腹部朝上躺在台盆边，但这张照片引起了高女士的怀疑。带着疑惑，高女士让客户再拍一段视频。 据齐鲁晚报报道，此后高女士选择发布视频曝光此事。 该视频曾被举报下架，但在修正后重新发布。高女士表示，自己主动修正，重新发布一条与原视频差不多的视频，隐去了顾客的隐私。"我本来不较真 的，让他吃就吃吧。 江苏苏州相城区大闸蟹经销商高晶：第一眼看上去就感觉像AI制作的，运镜很奇怪。如果再细看，螃蟹脚出现了 ...

每经编辑｜何小桃 据央视新闻报道，眼下，AI技术正赋能千行百业，为人们的工作、学习、生活带来很多便利。但最近，一些人竟然用AI制作虚假图片和视频向网店商家 索赔。来看江苏苏州一位大闸蟹经销商的遭遇。 买家发来"死螃蟹"图片视频 高女士对此也非常疑惑，立即联系消费者，请他们提供相应的证据。 很快，买家就发来了一张照片，画面显示，6只螃蟹腹部朝上躺在台盆边，但这张照片引起了高女士的怀疑。带着疑惑，高女士让客户再拍一段视频。 江苏苏州相城区大闸蟹经销商高晶：第一眼看上去就感觉像AI制作的，运镜很奇怪。如果再细看，螃蟹脚出现了5只小腿，和正常的大闸蟹不一样。视频 比例也裁切了，很明显就是为了去掉水印裁切的。所以我们让他再拍一段视频，用手指逐个去指一下，拨动一下，这个我觉得不可能再造假了。后来他真 的发来了这样一段视频。 疑点重重 11月17日，在江苏苏州相城区经营大闸蟹生意的高女士，根据客户订单，发出了一箱8只装大闸蟹。第二天，广州的买家就收到了货，完成签收。就在签 收后10多分钟，高女士收到平台发来的客户反馈。 江苏苏州相城区大闸蟹经销商高晶：11月18日早上11点40分左右就签收了，11点53分就来反馈说8只螃蟹 ...

继"网购8只螃蟹称死6只，疑似AI合成造假者被行政拘留8日"的新闻引发热议后，近日又有商家发布视频称自己售卖的行李箱遭遇了"AI造假+实物调包"的 类似经历——消费者收货后发来疑似AI生成的商品损坏图片要求退款，被商家要求退货后，竟调包寄回劣质替代品，而电商平台在商家未签收包裹的情 况下，直接将货款退还消费者。12月4日，扬子晚报/紫牛新闻记者从视频发布者陈先生处了解到，目前他已向警方报案。 近几个月以来，伴随"双十一"的消费热潮，有不少商家反映，部分消费者利用AI技术伪造快递损坏图、商品瑕疵图，以此申请"仅退款"的情况屡见不鲜， 给商家正常经营带来困扰。记者在实测中发现，如今的AI工具操作门槛极低，不仅能快速生成以假乱真的商品瑕疵图片，还能根据瑕疵图片精准生成对 应的视频，成品真假难辨。对此，律师表示，利用AI伪造商品毁损图/视频、虚构商品破损事实申请"仅退款"，若以非法占有为目的，虚构事实、隐瞒真 相骗取商家货款，可能涉及违法。 行李箱收货即发"破损图"要求仅退款？ 商家拒退遭调包，平台已退款 根据商家陈先生回忆，11月26日，一名消费者在其店铺下单了一款行李箱，约500元出头，11月27日收到货物。随后 ...

Core Viewpoint - The rise of AI-generated marketing videos has led to concerns about authenticity and consumer trust, as many of these videos blur the line between reality and fabrication, posing risks to consumer rights and safety [1][3]. Group 1: AI Technology in Marketing - The use of AI technology for mass-producing marketing videos is becoming increasingly common in e-commerce, with tutorials available online for creating eye-catching content [2][3]. - Current AI video generation models struggle with accurately depicting complex physical interactions, leading to issues such as "穿模" (body penetration), which highlights the limitations of AI in understanding real-world physics [2]. Group 2: Misuse of AI and Consumer Protection - There have been instances of individuals and brands being impersonated in AI-generated content, misleading consumers and infringing on their rights [4][5]. - Regulatory bodies are taking action against companies that misuse AI for false advertising, as seen in a case where a company was penalized for promoting a product using a fabricated video of a well-known media personality [5][6]. Group 3: Regulatory Responses and Industry Standards - Authorities are advocating for stronger regulations and collaborative efforts to address the challenges posed by AI in advertising, emphasizing the need for improved identification and management of AI-generated content [6][7]. - Platforms are evolving from manual reviews to AI-assisted identification of violations, enhancing their ability to detect and manage misleading content [7]. Group 4: Consumer Awareness and Reporting - Consumers are encouraged to report suspected AI-related false advertising through official channels, highlighting the importance of vigilance in maintaining market integrity [8].

Core Viewpoint - The recent incident involving an actor facing "AI impersonation" has sparked renewed public discussion about the implications of artificial intelligence, particularly in the context of content generation and potential misuse [1][2]. Group 1: AI Misuse and Public Concerns - The rapid development of generative AI has made video production accessible without specialized skills, leading to misuse such as fake buyer reviews and fraudulent content targeting vulnerable populations [1]. - The incident serves as a warning about the dangers of AI being used as a tool for deception rather than creativity and efficiency [1]. Group 2: Regulatory Measures - The "Artificial Intelligence Generated Synthetic Content Identification Measures," effective from September, mandates explicit and implicit labeling of AI-generated content to help users identify misleading information [1][2]. - Despite the implementation of these measures, some AI content remains unmarked, misleading audiences and necessitating a more robust governance framework [2]. Group 3: Recommendations for Governance - A multi-layered governance system is essential to combat AI-related fraud, including clearer legal standards for penalties, defined responsibilities among service providers, platforms, and users, and enhanced regulatory efforts [2]. - Upgrading technical capabilities for high-precision detection of fraudulent content is crucial for effective identification and mitigation of AI-generated deception [2].

Core Points - The rise of fraudulent refund claims in the e-commerce sector is causing significant distress for honest merchants, as some buyers exploit AI tools to create fake evidence of product damage [1][2] - The misuse of AI technology undermines the original intent of the "refund only" policy, which aims to enhance consumer experience while protecting merchants' rights [2] Group 1 - The fraudulent practices involve manipulating product images to appear damaged or spoiled, affecting a range of low-cost items, leading to frequent but small-scale financial losses for merchants [1] - Merchants face a dilemma where the cost of defending against these claims often exceeds the value of the products involved, highlighting the challenges in the current e-commerce landscape [1][2] - The introduction of new regulations in September aims to clarify the use of AI-generated content, prohibiting malicious alterations and ensuring that merchants can protect their rights [2] Group 2 - E-commerce platforms are urged to enhance AI image recognition technology and implement stricter review mechanisms to combat the rise of AI-assisted fraud [2] - As disputes over "refund only" policies continue to increase, many merchants are adjusting their after-sales strategies to better navigate the challenges posed by AI misuse [2]

Core Viewpoint - The incident involving actress Wen Zhengrong highlights the urgent need for legal and regulatory intervention regarding the unauthorized use of AI-generated images for commercial purposes, raising questions about platform accountability [2][4][6] Group 1: Celebrity Rights Protection - Celebrities must follow a structured approach to protect their rights, starting with evidence collection, such as saving screenshots of AI broadcasts and links to infringing products [2][3] - Legal actions can be taken against merchants for infringing on portrait rights and name rights, with the Civil Code providing a solid legal basis for such claims [3][4] Group 2: Platform Responsibilities - Platforms cannot evade responsibility and must implement preemptive measures, such as using technology to identify AI-generated content and verifying identities in live broadcasts [4][6] - Upon receiving reports of infringement, platforms are required to act within 24 hours to remove infringing content, as stipulated by the E-commerce Law [4][6] Group 3: Legal Framework and Enforcement - The Civil Code and E-commerce Law provide a clear legal framework for rights holders to notify platforms and enforce their rights against unauthorized use of AI [4][5] - Regulatory bodies need to increase penalties for violations, as demonstrated by past cases where companies were fined for impersonating public figures [5][6] Group 4: Challenges and Solutions - The covert nature of AI fraud complicates enforcement, but proactive monitoring and technological upgrades are essential for platforms to prevent misuse [5][6] - Collective action among celebrities, platforms, and regulatory authorities is necessary to effectively combat the misuse of AI technology [6]