AI short drama accused of facial recognition infringement involving celebrities
A television-broadcast AI short drama in China reportedly used AI-generated faces that blended characteristics resembling two well-known entertainers, sparking online accusations of portrait rights infringement and raising questions about how to identify and regulate AI facial mimicry in commercial content.
What happened
On 27 April 2026, the short drama 《中国传说·白蛇》 (The Legend of the White Snake) broadcast on Zhejiang Satellite Television drew criticism for reportedly using AI-generated facial features resembling entertainers Xiao Zhan and Wang Yibo. Network users noted that promotional materials featured male characters with facial resemblance to the two celebrities.
The production company 杭州九紫源文化科技有限公司 promoted it as China's first prime-time AI short drama on 14 February 2026. The drama aired on 17 February 2026 across multiple platforms and on Zhejiang Satellite Television during prime time, achieving second-place national viewership among 34 satellite channels in its time slot and exceeding 20 million total topic exposures online.
Reported impact
Legal experts noted that AI-generated faces in short dramas raise difficult questions of identification and infringement standards. A lawyer stated that if a general audience can identify an AI character as a specific person, infringement may be established regardless of whether the resemblance is intentional or accidental.
Index analysis
The incident reflects broader challenges in AI short drama regulation. Experts identified four types of portrait infringement: completely AI-generated faces, directed AI generation guided by keywords, direct AI face-swapping, and blended faces combining multiple people's features. This case appears to involve blended facial characteristics, making liability and infringement determination complex.
Governance controls
The production company registered a copyright for the work in February 2026 as a visual art piece. Industry experts recommend platforms establish categorical audit mechanisms, build feature databases for public figures to enable proactive detection, and simplify complaint procedures for portrait rights holders.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.