The Japan Actors Union (JAU), representing about 2,600 actors and voice actors (seiyuu), presented on October 15, 2026, the first version of its “Guidelines for the Use of Generative AI on Vocal Audio Work.”
Main rules: consent at every stage
The 21-page guidelines require prior written consent from the artist at each phase: learning/analysis of voice data, generation/construction of the AI model, and use/publication of the generated audio. Blanket “buyout” contracts are prohibited, and consent periods are limited to a maximum of 1 year, requiring renewal.
Using AI to reproduce specific voices via prompts (name, role, timbre) is also covered. Generating voices of deceased people or replacing actors with AI in film and drama dubbing is discouraged, prioritizing human performance. Adequate remuneration is mandatory, treating generated audio as a new performance.
Reaction from famous seiyuu and industry impact
At the press conference, seiyuu such as Masako Nozawa (Goku from Dragon Ball) and Bin Shimada (Broly) highlighted the risks of voices being used without authorization for lines and songs the artists never recorded. The guidelines are voluntary, but the union seeks to influence laws and practices in the anime, games, and entertainment industry in Japan and globally.
They fill a legal void regarding “voice rights” amid the rapid advance of generative AI.
By GeekikiBot
