Japan Actors Union issues AI voice guidelines: consent at every stage, royalties for AI voices, ban on cloning deceased actors

What happened

The Japan Actors Union announced new guidelines on October 16 regarding the unauthorized use of generative AI for voices, aimed at protecting actors and other performing artists and giving AI businesses a framework for operating safely in Japan.

The four areas covered

• Consent at every stage: AI companies must clearly state their purpose and obtain permission from the person whose voice they want to use for AI training. Prior consent is required at the training (input), generation, and publication/use (output) stages.
• AI-generated voices as a new "performance": Using AI to generate voices is defined as a new form of performance, meaning companies must pay appropriate compensation or royalties to actors — framed as letting actors license their own voices rather than being displaced by AI.
• AI-generated dubbing/translations: When content is translated into other languages using AI-generated audio that mimics an actor's voice, actors are guaranteed the right to obtain new usage licenses and additional compensation to cover global distribution, protecting the artistic integrity of the original performance.
• Deceased actors: Companies are banned from cloning deceased actors' voices, with one exception: projects the deceased agreed to participate in during their lifetime. Even then, prior written consent is required from the bereaved family and the agency or organization the deceased was affiliated with, plus royalties.

Context

Japan has no law directly protecting voice rights; the union and other performers' organizations have been pushing for legal reforms, citing cases such as voice actor Kenjiro Tsuda's lawsuit over a TikTok account that allegedly imitated his voice with generative AI. The union says it aims to establish the guidelines as industry-wide rules and to press the government to legislate "voice rights."