Chinese actress Wen Zhengrong recently found herself trapped in a surreal “battle for identity.” While browsing live streams late at night, she was stunned to discover more than 30 “versions” of herself appearing simultaneously across different platforms — switching between red and white outfits, delivering eerily similar sales pitches. These AI-generated “digital clones” not only had mismatched lip movements and robotic expressions, but still managed to draw in viewers with their highly realistic likeness. The absurdity peaked when Wen confronted one in the comments, asking, “If you’re Wen Zhengrong, then who am I?” — only to be instantly blocked, in a real-life parody of “the imposter silencing the original.”

演员温峥嵘近日陷入一场荒诞的“身份保卫战”。当她深夜刷直播时，竟目睹超过30个“自己”在不同平台同步带货，红衣白衣轮换登场，推销话术高度雷同。这些AI合成的“数字分身”不仅口型对不上声音，表情更是僵硬如提线木偶，却仍凭借高度相似的形象收割流量。更讽刺的是，当温峥嵘在评论区发出灵魂质问“你是温峥嵘，那我是谁”时，得到的回应竟是被光速拉黑，上演了一出现实版“李鬼拉黑李逵”。

According to Wen Zhengrong’s team, since April 2025, fake accounts have been multiplying endlessly — “cut one, and two more appear.” Fans once set a record by reporting 18 fake accounts in a single day, yet the cycle persists: the speed of bans can’t keep up with the speed of rebirth. On a secondhand trading platform, “face recognition” and “video production” services are openly sold for as little as 30 yuan, with sellers even “kindly reminding” buyers to ensure their materials “do not infringe copyrights.”

据温峥嵘团队透露，自2025年4月起，假账号便呈现“割一茬长一茬”的态势，粉丝曾创下单日举报18个假账号的记录，却始终陷入“封禁速度赶不上再生速度”的怪圈。某二手交易平台更以30元的价格公开售卖“人脸识别”“视频代做”服务，卖家甚至贴心提示“需确保素材不侵权”。



Although the Administrative Measures for Labeling AI-Generated Content, implemented in September 2025, require all synthetic content to be clearly labeled, merchants easily evade supervision through tricks like “mirror flipping” or “picture-in-picture masking.” Some even hide disclaimers deep in the interactive interface. For victims, the cost of a single facial likeness authentication can reach 8,000 yuan, and tracing cross-platform infringement chains is so complex that most simply give up.

尽管2025年9月施行的《AI生成内容标识办法》明确要求合成内容显著标注，但商家通过“镜像翻转”“画中画遮挡”等技术轻松规避监管，更有甚者将标识藏在交互栏角落。对受害者而言，单次肖像权鉴定费高达8000元，而跨平台追踪侵权链条的复杂度，更让多数人望而却步。



On November 6, Douyin Group Vice President Li Liang responded that, although investigations confirmed that the case of “three livestreams running simultaneously” did not occur on their platform, they did uncover instances of AI-generated celebrity impersonations in short videos.

11月6日，抖音集团副总裁李亮回应称，虽核查确认“三个直播间同时开播”案例未发生在本平台，但已发现部分短视频存在AI仿冒名人带货行为。

Douyin E-commerce has since launched a special crackdown — since October, it has dealt with 11,000 infringing influencer accounts, removed over 6,700 counterfeit items, and encouraged creators to activate the “block similar infringing content” function. The platform emphasized that AI applications must remain compliant, warning that any misuse or abuse will face severe penalties, including account suspension, product takedowns, and frozen transactions.

抖音电商随即启动专项治理，10月以来处置侵权达人账号1.1万个，下架仿冒商品6700余个，并引导达人开通“阻止相同侵权内容传播”功能。平台强调，AI应用必须合规，任何误用、滥用行为都将面临账号中断、商品下架、成交冻结等严厉处罚。

本文来自公众微信号：听歌学英语