AI could have 'additive' role for audiobooks but consumer response unclear, LBF hears

Audiobook publishers and retailers believe AI has an “additive” role to play with audiobooks, especially around discovery and production in foreign languages and smaller markets, but say it remains to be seen how consumers will react to increased use of synthetic and digital voices.  

Speaking on the panel "The Future of Audio in Publishing: Global Trends and the Impact of AI" at London Book Fair yesterday (13th March), Michele Cobb, executive director of the Audio Publishers Association, told delegates that foreign-language audiobooks from American and British publishers were a “quickly growing” part of the market and “some of those are going to come forward not only with a human voice, but also with an AI voice”. 

She said: “That’s something that’s starting to infiltrate the market and we’re trying to understand as a group how the consumer feels about that. Because publishers, and retailers of course, ultimately want to think about the consumers."

Amanda D’Acierno, president and publisher at Penguin Random House Audio, argued that the English-language market had grown so much due to “great production”. “Whether it was Audible, or a Big Five competitor, we were all really aligned in making the best possible productions out there.” She continued: “As we lean into non-English language we want to have that same great experience there. Production has to be something that people want to listen to and pass on.” 

Aurelie de Troyer, Audible’s senior vice-president and head of content for UK & Europe, agreed, explaining that original content was a “big priority” for Audible, which also adds a lot of multimedia, interviews and soundscaping to its audiobooks to “reinvent ways to deliver books to people”. However, she stressed: “They have to be high quality.”

Panel chair Videl Bar-Kar, vice-president for audio at Bookwire GmbH, said AI in audiobooks starts in production, helping to check for errors, but noted that people felt more cautious about the idea of synthetic voices taking over jobs from humans.

On day one of the fair, Simon & Schuster US president and chief executive Jonathan Karp said the publisher was going to "experiment" with AI narration "in languages where authors of old books might not otherwise ever be published".

Bar-Kar argued that it was "important to let consumers decide at the end of the day”, noting: “There might be business cases and use cases, especially in certain markets, where you really don’t have enough to get a minimum catalogue to get anyone to really invest in marketing a service and getting consumers into audiobooks you need”. He stressed that AI should have an “additive” role with audiobooks. 

An example is Storytel’s Voice Switcher, which launched in Poland in October and Sweden in February and allows users to select different voices to tailor their listening experience. Helena Gustafsson, chief content officer, told audiences that a recent survey of users showed that 89% of customers had at one point deselected a book because they didn’t like the human voice. “Voice Switcher is not replacing the human narrator [...] you also then have an addition to choose between three or four synthetic voices that we have developed,” she said. 

Owen Smith, vice-president of product and technology for audiobooks at Spotify, said it was important to give consumers, as well as authors, the option of digital and synthetic voices. "There are so many authors that can’t publish their book due to the barriers of entry and the cost of production,” he said.  

He discussed Spotify’s use of AI on the discovery side, using algorithms to match users with the right book, especially as the catalogue grows. This mimics what the Swedish tech giant has done with music and podcasts. “It’s a perfect way into the format for listeners who may never have considered an audiobook before,” he said.