Kenjiro Tsuda sues TikTok over unauthorized AI voice cloning videos

This litigation establishes a critical precedent for how social video platforms must moderate synthetic media that infringes on publicity rights. If the court holds TikTok responsible for failing to remove AI-generated clones, it could force global platforms to implement more aggressive automated detection for vocal deepfakes. The outcome will likely influence how streaming services and social apps manage talent likeness in the generative AI era, particularly in markets like Japan where voice acting is a central pillar of the media economy. Watch for the Tokyo District Court's verdict this Wednesday to see if vocal identity receives the same legal protections as physical likeness.

The Tsuda v. TikTok case is the first in Japan to test whether a platform bears legal responsibility for hosting AI-generated content that clones a performer's voice. According to court records accessed by AFP, TikTok argued it merely hosted a forum where users could post creative content within the scope of free speech, disputing that the voice in question was anything more than a generic male baritone. The offending account, which used a profile image resembling a Jujutsu Kaisen character voiced by Tsuda, had accumulated more than 210,000 followers and generated between 500,000 and 750,000 yen in monthly revenue from view counts, according to the complaint. The Japan Actors Union, through executive director Yuko Sasaki, has publicly backed Tsuda and launched a campaign called "No More" against unauthorized generative AI use of performers' voices.

On the regulatory side, Japan's Justice Ministry published non-binding guidelines in August 2026 clarifying that a person's voice qualifies as a symbol of personality and falls under existing publicity rights and portrait protections. The ministry had convened an expert panel in April 2026 specifically to address the rise in unauthorized AI-generated voice use, as NHK World reported when the panel began discussing what cases could be subject to civil liability. These guidelines are not legally binding, and it remains unclear whether the Tokyo District Court will reference them in its ruling. The International Federation of Actors has separately noted that copyright and data protection laws in many countries have not kept pace with AI cloning threats to performers.

From a technical and industry perspective, the Tsuda case highlights a gap in content moderation tooling for vocal identity. TikTok's defense rests on the claim that the voice is generic and not identifiable as Tsuda's, yet Jiji Press reported that audio analysis of the videos found the synthetic voice closely resembles Tsuda's distinctive deep and mellow timbre. Waseda University intellectual property professor Tatsuhiro Ueno told AFP it is very unlikely the court will deny Tsuda's publicity rights outright, and he expects the verdict to draw international attention given the actor's recognition in European markets through Jujutsu Kaisen. The case was originally filed in November 2025, with at least 188 infringing videos identified between July 2024 and September 2025, making it a benchmark for how platforms must handle takedown requests involving synthetic vocal media.