Artificial intelligence is beginning to change the world of audiobooks, with the ability to create sequential recordings without each time using a human to take over the task of narration with the voice, in a development that worries workers in the field of audio recording that has been shrinking for years.
Tanya Ebe has been recording scripts with her own voice for 20 years, doing the job full time. As of 6 months ago, her activities have halved, as have many of her colleagues whose work has slowed.
“It is logical that artificial intelligence affects us,” she says, and “I think the future plan is based on replacing employees (with artificial intelligence systems) to reduce costs.”
While there is currently no announced format for audiobooks developed with artificial intelligence, experts in the sector confirm that thousands of these books developed based on an audio database are now on the market.
by “a quarter of the value”
Among the most advanced services in this field, Deep Zen provides a technology through which an audiobook can be developed at a quarter of the value of traditional projects.
The small London-based company relies on a database it created by recording the voices of different actors who were asked to express a variety of emotions.
“We signed a licensing agreement with all the sounds we use,” says Kamis Taylan, general manager of Deep Zen. “We pay for the recordings in addition to paying royalties every time we use a sound for a project.”
“Things are moving so fast that many of the newcomers don’t have the same ethics,” warns Tanya Ibe, who use unpaid votes.
Kamis Taylan explains that “there is a gray area” that many platforms exploit, as “they take different voices for you, me, and five other people, to make a new one without paying anything, under the pretext that the voice does not belong to anyone.”
All companies contacted by “Agence France Presse” denied carrying out such practices.
In addition to its pre-recorded voices, the startup “Speechky” allows the use of existing voices in existing databases, according to General Manager Dima Abramov, who explains that this second feature requires the signing of a contract stating usage rights.
The five major roles operating in the sector in the US market did not respond to the agency’s questions on this matter.
But according to professionals, many names in the traditional publishing field are already using so-called generative artificial intelligence, a technology capable of creating texts, images, videos or sounds, without human intervention, based on existing content.
“Professional storytelling has always been, and will continue to be, essential to listening to Audible,” said a spokeswoman for audiobook giant Audible, a subsidiary of Amazon.com.
“However, as technology improves, we envision a future where human interpretation and (using AI)-generated content can coexist,” she added.
The tech giants are investing heavily in artificial intelligence and participating in the booming economy of audiobooks created with this software.
At the beginning of the year, Apple launched an offering that includes “digital storytelling,” which, according to the company, aims to “make audiobook creation accessible to everyone,” especially “authors, freelancers, and small publishers.”
Google offers a similar service, which it describes as “self-narrative”. “We need to popularize the publishing industry on a large scale, because in the current situation, only the most famous names are turning into audiobooks,” says Kamis Telan.
“Automated storytelling opens the door to all existing books that were not recorded (with audio), and for all future books that would never have been converted (into audiobooks) due to economic restrictions,” explains Dima Abramov, noting that 5% of the books that currently exist are converted. audio books.”
He notes that the growth of the market will benefit voice actors, as “they will make more recordings and earn more money.”
“Storytelling, at its core, is allowing humans to reconnect with their humanity,” says Emily Ellet, president of the Association of Professional Audiobook Tellers, arguing that “storytelling must remain entirely human.”
For Tanja Ibe, AI storytelling “lacks emotional connection. There’s a real difference (with classical recordings). But over time, listeners might get used to it. I think it does.”