Amazon will begin inviting a small group of Audible voice artists this week to train AI-generated voice clones of themselves, with the goal of speeding up audiobook production for the platform. The U.S.-only beta test was announced on Audible's Creator Marketplace and will expand to rights holders such as authors, agents and publishers “later this year,” according to Amazon.
“There is a huge catalog of books that don't yet exist in audio, and as we explore ways to bring more books to life on Audible, we are committed to carefully balancing the interests of authors, speakers, publishers and listeners,” Amazon said in its announcement. Participants in the beta will submit a voice recording to train their AI replica, and will retain control over the projects they want to audition for, both in live performances and AI-generated recordings.
Voice actors can also use Amazon's production tools to edit the pronunciation and pacing of their AI voice emulation if a rights holder selects them for a project, and they can review the final production for errors or inaccuracies. Amazon says voice actors will be compensated via a “royalty share” model on a “title-by-title basis,” but has not elaborated on how much voice actors can expect to earn.
“Audible will not separately use a speaker’s voice emulation for any content without their consent.”
The announcement blog states that beta participants can create a voice emulation “for free,” meaning there may be an upfront cost for narrators in the future when the feature becomes generally available. All titles read with voice emulations will be labeled on the product detail page. “Narrators control which works are read with their voice emulation,” Amazon said. “Audible will not separately use a narrator's voice emulation for content without their consent.”