Leading audiobook executives from across the publishing industry gathered at the London Book Fair earlier this month for a panel addressing how artificial intelligence is reshaping the audiobook market. Many expressed optimism about the potential for growth in the space presented by AI technologies—despite lingering concerns over rights management and compensation models.

"We're seeing AI affect every area of the audiobook business," said Ana Maria Allessi, president and publisher of Hachette Audio. "And we're at the very early stages of testing and developing and trying to be responsive to our authors in supporting what they think they want to do or the questions that they have around AI voice."

Contrary to what some might expect, "one might think that there's been a consistently negative response to the technology, but that's actually not the case," Alessi explained. "There are many authors who have come to us asking very distinct questions—and asking how we can support them using AI voice."

Navigating new realities

At first, AI narration might appear to offer publishers an opportunity to quickly convert backlist titles into audiobooks. But Allessi pushed back against this assumption.

"There's this impression that the Big Five publishers have this vast backlist that has not yet been exploited, and that's available to us. And if only we had an AI voice, we could do more," Allessi said. "But every one of us has spent a considerable amount of time reviewing our backlist, and most of those titles have already either been exploited directly by us, or they've been sub-licensed to the many audio-only publishing houses."

Swedish e-book and audiobook subscription service Storytel has been experimenting with AI through what CEO Helena Gustafsson called "the voice switcher feature," which allows users to select from several AI voices when listening to content. "We did it this way because we wanted to do it respectfully towards the narrators," Gustafsson explained. "But it was also to understand and learn how people actually select voices when they have the opportunities."

When asked about consumer behavior regarding AI voices, Gustafsson noted that "the listening audience for AI voices is still very small. If you have a choice and you can choose a human narrator, you do that." However, she added, cloned voices from popular narrators see significant usage: "We have some of our great voices that are licensed. They are cloned from super popular narrators, and those get a big amount of listens."

The panel also addressed the delicate issue of narrator rights and compensation. Allessi emphasized that there's "a big difference between a cloned voice and a synthetic voice," and highlighted the importance of working with narrators on terms and conditions.

"We've tried to be very communicative and attentive and supportive of the narrator community to try to work this out the right way," Allessi said. She recounted a conversation with representatives from SAG-AFTRA about voice cloning: "It was very interesting to me that they were open to discussing it and they saw this as a very nuanced issue. This was not 'hair on fire, and no way you're going to ruin our business.' It was, 'Oh, what circumstances? What are the terms?' "

Allessi noted that while AI can more easily facilitate many things, such as the automated creation of multi-cast/mult-voice productions, the question remains whether or not "the customers actually want that sort of thing."

Outlook and opportunities

Storytel has ventured into more experimental territory of late with an AI novelist project. Gustafsson described working with "a fictitious author" named Rosy Lett, an anagram of the developer's name. The "author" is an AI creation, appearing as the avatar of a 22-year-old mixed race female author. Lett's book, New Horizon, is a collection of short stories that explores the intersection of humanity and artificial intelligence. The four tales are set in the fictional burgh of New Horizon City, in a future that finds AI deeply integrated into society and everyday life.

Upon the project's release in January, Gustaffson hosted a press conference with Lett’s avatar on Instagram Live, taking questions— some of them pre-prepared by Storytel's team, with answers also prepped in advance, as well as questions from the Instagram audience, with answers generated by AI. Gustaffson did not indicate if the book was a success with readers or listeners, but did say that she would probably not publish a follow-up at this time.

Eric Bartoletti, head of business development at Bookwire, said that his company has received interest from tech firms seeking to license audiobook content for AI training—particularly for languages that haven't yet been a primary focus, such as Portuguese. "The things that we're currently discussing with potential partners are about the voice," Bartoletti said. "They usually want to get the text in addition to the audiobook, but it's not so much about using the text to put in a foundational model and enhance it. It's really about training the model to make the AI voices sound more human."

Panel members expressed optimism about the future of AI technology in the audiobook business, but acknowledged that business models remain uncertain. Allessi emphasized the need for standardization and equity in AI implementation, while Gustafsson suggested that the industry needs to be proactive, asserting that "if we will not make decisions on how we will use AI, someone else will do it.”

For his part, Bartoletti noted a significant shift in industry attitudes. "Looking back at last year, I saw a lot of skepticism or even fright before publishers regarding AI. They did not really know what to do with it, and they thought, 'Okay, the AI is taking over, and we can go out of business,' " he said. "Six or seven months later, the vibe was completely different. Most of the people—actually all of the people I'm talking to—focus on the opportunity."