Amazon's Audible has launched a groundbreaking beta program allowing narrators to create AI-generated voice replicas of themselves, potentially revolutionizing audiobook production and raising questions about the future of human narration in the industry.
AI voice generators have been on the rise lately as a lot of creators and podcasters make use of them to streamline their tasks and enhance their posts12. With all the hype, Amazon has decided to let Audible join the fray3. Unlike most AI audio tools, Audible's voice cloner feature isn't designed for general use. Since Audible is an audiobook platform, the purpose of its AI is specific: allow audiobook narrators to clone their voices so authors on the platform can use the voices they like for their audiobooks34. Of course, the voice actors will be compensated fairly for their AI-replicated voices, and like any audiobook narration gig, they can choose which projects they want to accept35.
Audiobook narration is a long, delicate process. The voice actors should have the skill and patience to read books from cover to cover, reciting every word and every chapter as flawlessly as they can. This is not an easy feat as a single book can have hundreds upon hundreds of pages.12 Narrators should also be particular with a lot of factors— how fast their reading pace is, how they must portray the characters, which tone they should speak in, whether the readers can clearly understand their pronunciation, and other essential requirements.2 With AI voice cloning available in Audible, they can earn their keep without having to go through that same process again and again. After they're able to record their voice once, AI can do the work for them by creating the narration for an entire novel. They can monetize their talents in less amount of time and without repetitive effort.13
Authors can also benefit from this feature, especially if they don't have enough budget to spare for an audiobook narrator. Because the voices are cloned and not the real deal, they should be priced cheaper, making them a cost-effective alternative12. The voices are still modeled after the audio recordings of actual narrators so the quality won't be that far off from the original3. If an author is also rushing to get their audiobook out, AI can make the narration process faster so they can receive the finished output right before the scheduled launch4. Another advantage is that authors can simply browse through a library of voices, allowing them to listen and compare the options until they finally find the right match for their book5.
Right now, Audible's voice cloner is still in its testing phase. The audiobook service only gave access to a few select narrators who are willing to have their voices cloned for the platform1. Audible posted on its channel, ACX (Audiobook Creation Exchange): "We're taking measured steps to test new technologies to help expand our catalog and this week we are inviting a small group of narrators to participate in a US-only beta enabling them to create and monetize replicas of their own voices using AI-generated speech technology."1 They also said that they were testing the feature first to see the results and make corrections to the system in case there were errors such as mispronunciations1.
If you're an author, the answer depends on your needs and resources for your audiobook. When you have the money to hire an audiobook narrator, then find one. Real humans are still the better choice if you have the budget since these professionals can inject the right emotions and inflections into their words, which AI can't replicate1. Sure, AI may be able to recreate the voice, but a narrator's way of speaking is their own. Their ability to fit their voice to a particular scene or a certain context is still unattainable with the current AI audio technology2. However, if you're still starting out in the publishing industry, you may not be able to afford your ideal audiobook narrator. Audible's AI voice cloning function gives you the chance to still get a reliable narration of your book at a more affordable price3. You can pick the voice of the narrator you prefer, and you will have an audiobook you can proudly release to eager readers4.
Integrating AI voice cloning in the book publishing sector is a smart move on Audible's part. Narrators can earn from less effort, while authors can get quality audiobooks at less cost. It's a win-win situation for both parties.1
It must be noted, though, that despite Audible's announcement, the submission requirements on ACX still state that they will only allow human-narrated audiobooks. This is still subject to change as Audible hasn't launched their AI voice cloner to the public yet. We suspect that once they finish the development of this feature, those standards will be revised to fit the new offering. They may allow AI-generated narrations from their own platform but maintain the ban on replicated audio and AI voices from other sources.2
For now, the Audible team and the narrators they have onboarded are doing their best to make sure that the resulting audiobooks from cloned voices come out great and free of mistakes. The AI is still in its beta stage, so it's open to further improvements. Authors and audiobook narrators alike should keep an eye out for the major release if they want to try voice cloning for their work.3