top of page
  • Writer's pictureThe Publishing Post

HarperCollins: Audiobooks and AI

By Cameron Phillips, Kathryn Alley and Rose Cook

Earlier this month, it was announced that HarperCollins would be partnering with audio AI company ElevenLabs to create audiobooks for its foreign language business. With AI a contentious topic across the publishing industry, what will this mean for the future? The audiobooks team share their thoughts.


When reading the headlines about HarperCollins and audio AI, I was initially intrigued. Upon further reading, however, I began to feel more conflicted about this move and its possible implications.

The Bookseller stated that ElevenLabs’ goal with developing this technology is to “make spoken content universally accessible in any language and in any voice.” Enabling greater access to titles, including foreign language titles, surely cannot be a bad thing. For an industry trying to improve the diversity of what they publish, AI which can create audiobooks in multiple languages would be an invaluable tool. 

However, using AI to create an audiobook usually developed by a team in collaboration, and who cast, record and edit the work, leaves me with some concerns. Will the quality of the production be maintained? Can software replicate the intonation and emotion of the human voice? What implication does this have for jobs if the work of many is done by a computer? 

Answers to these questions will only become clear with time. Whilst still conflicted, I remain optimistic about the possibilities for accessibility and equality within audio publishing that AI could bring, potentially making new books accessible to audiences that would not have been able to listen to them before. I look forward to listening to an AI audiobook for the first time and to experiencing this technology and developing my thoughts.


This is not the first time we have discussed audiobooks and AI, yet here we are again. I think one of the things that strikes me from this fresh debate regarding the medium is the line from ElevenLabs’ CEO, Mati Staniszewski: “Without AI, there simply wouldn’t be the time or resources to make this happen.” Now, this feels to me like someone concerned with targets and margins, which no doubt are key to publishing. However, on the other hand, people who work in publishing do so because they love art, literature and the creative space, and this seems a far cry from those ideals.  

Any form of AI technology in publishing, especially in audiobooks, should not come at the cost of human emotivity, as that is what breathes life into the words on the page and the performances of these words in the studio. Undoubtedly, businesses are commercially driven, so margins and efficiency must be considered, but those should not come at the cost of quality, with a large swathe of audiobooks pushed out without proper care and treatment. There are so many brilliant writers and vocal performers out there, and if it takes a little longer to find them, a little longer to complete a project and a little more money, then, surely, so be it. 


Like Rose, I was initially intrigued and somewhat hopeful when it was announced that HarperCollins would be partnering with ElevenLabs to utilise audio AI for a selection of backlist titles. For many authors, the challenge lies not in choosing between a human narrator and an AI narrator, but between an AI narrator and not having their book narrated in audio at all. ElevenLabs CEO Staniszewski writes, “Now, every author can see their work come to life in audio, readers can be offered more choice, and the linguistic barriers of content can be dissolved.” A partnership with an audio company like ElevenLabs would make audio narration a reality for many stories that have never become celebrated listens. 

However, it’s concerning to me that a process that would take a team of qualified individuals much time and effort to complete is now anticipated to take AI an hour or less. The process is leaning more towards a user-driven service where AI voices are selected and pieced together within the listen. My main concern within the AI and audiobook controversy is how the quality of human narration can never be fully replicated or produced artificially. The listens I continue to think about long after hearing them are the stories where the narration resonated with me deeply. This kind of empathy and thoughtful intention is what I have found AI to seriously lack. 

Thankfully, HarperCollins announced that they will continue to devote time and resources to voice actor-focused productions, whereas the AI text-to-speech option will be more of a complementary tool. I look forward to exploring these new AI-narrated listens as they are produced and hope that they tell each author's story with as much emotion and consideration as they certainly deserve. 



bottom of page