One of the most prominent tools in the modern world that has been gaining traction in book publishing is Artificial Intelligence (AI). The term is used to refer to databases that enable problem-solving and can make predictions. Alan Turing, on the other hand, defined it as “systems that act like humans.” This definition is particularly relevant to book publishing, as this is an intrinsically human industry that deals in great part with creative content, not just quantifiable data.
In the case of book publishing AI is already being used to format texts automatically and by streamlining the copyediting process, and is helping to collect data from the audience in order to market a book more effectively. All of this has made it particularly useful for authors who want to self-publish. Yet there has also been controversy regarding its use to generate art for book covers, as it goes against the ownership authors have over their work in order to generate a new image.
On today’s blog we will be discussing one of its latest implementations in publishing, audiobook narrations using AI. Different players in the industry have started selling books narrated by an artificial or synthetic voice thanks to AI technology that can process a text and turn it into audio. Such is the case of Apple Books, which launched an AI tool aimed at self and small publishers that allows them to turn their books into audiobooks. Nevertheless, they can only be distributed through their platform. In a similar fashion, Google Books also offers a varied selection of voices that authors can pick from to narrate their books.
This seems like an important new application considering that audiobooks saw a 25 percent increase in sales last year according to The Guardian. It could potentially help self-publishing authors and small publishers increase their audiobook output by shortening their production times and reducing costs. It would help them gain presence in a market constantly dominated by big publishers that can easily pay voice actors, and celebrities, to narrate their audiobooks.
Still, there are detractors to this new technology that have aimed their darts from different directions. Some of its critics (as can be seen in the reviews available in the different platforms that offer AI narrated audiobooks) point to how unnatural or robotic some of them sound, which drives them away from the whole experience. While some platforms offer narrations that can mimic the human voices pretty accurately, there are still instances in which you can tell the uncanny valley effect of the narration.
This advent is also questioned by some alluding to the fact that it takes away value from the experience as it takes artistic expression away. This directly impacts voice actors, as the cheaper AI powered alternatives threaten to take away an important part of their business. Still, there is also potential for new business models that use an actor’s voice as a template and then pay them royalties for their use by an AI. It will be interesting to see how copyright laws will have to adapt in order to protect the rights of voice actors just as it is currently happening with art.
There is much still in the air regarding a widespread implementation of AI for audiobook production, and one can only speculate at the moment about the direction that the industry will take. So far its biggest contender is that the technology still has a certain level of uncanny valley to overcome in order to compete with the craft of voice actors. Only time will tell to what extent machines can truly behave like humans, as Turing described it, and provide a truly authentic narration.
Written by Francisco Cabre Vasquez.