3月20日，話題#登上熱搜。有網(wǎng)友稱AI短劇《重生后，我成了娘親的守護(hù)神》第14集開頭疑似使用了演員楊紫的臉。

Chinese actress Yang Zi has condemned the unauthorized use of her likeness in an AI-generated short drama, as a related case involving Dilraba Dilmurat underscores growing legal scrutiny over AI face-swapping in China.

3月20日，楊紫工作室發(fā)文：針對某作品未經(jīng)授權(quán)，利用AI換臉、深度合成等技術(shù)，擅自使用楊紫女士的肖像制作并傳播視頻內(nèi)容的行為，嚴(yán)重侵犯楊紫女士合法權(quán)益。工作室已委托律師對侵權(quán)內(nèi)容進(jìn)行取證，并將通過法律途徑追究到底！

目前相關(guān)視頻已下架。

On Friday, the topic trended online after netizens alleged that a short drama used AI technology to replicate Yang's appearance in one of its episodes. Her studio said the production had used AI face-swapping and deep synthesis without authorization, seriously infringing on Yang's rights, and added that legal action is underway. The video has since been removed.

迪麗熱巴被短劇AI換臉，法院判了

同日，演員迪麗熱巴起訴某平臺播出短劇利用AI換臉技術(shù)盜用其肖像牟利一事，也有了最新進(jìn)展。

據(jù)北京互聯(lián)網(wǎng)法院消息，法院認(rèn)定，短劇制作方擅自使用深度合成技術(shù)生成與某演員高度相似的形象，侵害其肖像權(quán)；短劇播出方未盡合理審查義務(wù)，同樣承擔(dān)相應(yīng)責(zé)任。

On the same day, a court ruling in a separate case involving Dilraba reinforced accountability for AI misuse. The Beijing Internet Court found a production company liable for infringing on an actor's portrait rights by using deep synthesis without consent. A distribution platform was also held responsible by the court for failing to conduct adequate review.

此案原告為國內(nèi)某知名演員，該演員發(fā)現(xiàn)被告A公司制作并發(fā)布的短劇中，通過AI換臉技術(shù)將其肖像拼接至劇中角色面部，相關(guān)話題在多個(gè)社交平臺引發(fā)討論，大量網(wǎng)絡(luò)用戶誤認(rèn)為原告參與了該劇。與此同時(shí)，被告B公司在其運(yùn)營的視頻賬號上線涉案短劇。

法院判決A公司、B公司分別在其運(yùn)營的涉案視頻賬號向原告發(fā)布書面致歉聲明并賠償經(jīng)濟(jì)損失。雙方均未上訴，一審判決已生效。

Both parties were ordered to issue public apologies and pay compensation, with the ruling taking effect after neither side filed an appeal.

部分平臺開始限制“AI換臉?biāo)嚾恕毙袨?/strong>

2月12日，豆包旗下視頻生成模型Seedance2.0全端上線后，執(zhí)行嚴(yán)格限制：禁止生成明星、公眾人物、網(wǎng)紅等真實(shí)人物肖像/視頻/語音。

查詢后豆包回復(fù)：2026年2月13日起全平臺（含豆包主App）統(tǒng)一收緊，關(guān)鍵詞攔截+人臉比對+內(nèi)容審核三重機(jī)制，命中即拒絕生成。核心原因?yàn)樾は駲?quán)、版權(quán)合規(guī)，防范AI換臉/偽造/商用侵權(quán)。限制范圍包括禁止生成明星姓名、肖像、聲音、視頻、換臉、模仿等；限制范圍不僅明星，所有未授權(quán)真實(shí)人物均受限，但用戶本人數(shù)字分身（需活體校驗(yàn)）不受限。

The controversy comes as platforms tighten rules on AI-generated content. Following the launch of Seedance 2.0, Doubao said it would ban the generation of real individuals' likenesses, voices and videos without authorization, using filters, facial recognition and content checks to curb deepfakes and infringement.

《法治日報(bào)》律師專家?guī)斐蓡T、北京市康達(dá)律師事務(wù)所合伙人羅娟表示，濫用AI換臉、聲音克隆主要涉及民法典第一千零一十八條（肖像權(quán)保護(hù)）、第一千零二十四條（名譽(yù)權(quán)保護(hù)）。未經(jīng)同意商用他人肖像、聲音，直接侵犯肖像權(quán)；若生成內(nèi)容虛假誤導(dǎo)（如虛假代言、不當(dāng)言論），會損害他人名譽(yù)權(quán)。

Legal experts cited provisions in China's Civil Code, noting that the unauthorized commercial use of a person's likeness or voice constitutes a violation of portrait rights, while misleading or fabricated content may also infringe on reputation rights.

來源：中國青年報(bào) @楊紫工作室 荔枝新聞 北京晚報(bào) 中國新聞網(wǎng) 九派新聞 法治日報(bào)

跟著China Daily

精讀英語新聞

“無痛”學(xué)英語，每天20分鐘就夠！