NEW YORK: As people brace for the disruptive impact of artificial intelligence on jobs and everyday living, those in the world of audio books say their field is already being transformed.

AI has the ability to create human-sounding recordings — at assembly-line speed — while bypassing at least part of the services of the human professionals who for years have made a living with their voices.

Many of them are already seeing a sharp drop off in business.

Tanya Eby has been a full-time voice actor and professional narrator for 20 years. She has a recording studio in her home.

But in the past six months, she has seen her workload fall by half. Her bookings now run only through June, while in a normal year, they would extend through August.

Many of her colleagues report similar declines.

While other factors could be at play, she said, “It seems to make sense that AI is affecting all of us.” There is no label identifying AI-assisted recordings as such, but professionals say thousands of audiobooks currently in circulation use ‘voices’ generated from a databank.

Among the most cutting-edge, DeepZen offers rates that can slash the cost of producing an audiobook to one-fourth, or less, that of a traditional project.

The small London-based company draws from a database it created by recording the voices of several actors who were asked to speak in a variety of emotional registers.

“Every voice that we are using, we sign a license agreement, and we pay for the recordings,” said DeepZen CEO Kamis Taylan.

For every project, he added, “We pay royalties based on the work that we do.” Not everyone respects that standard, said Eby.

“All these new companies are popping up who are not as ethical,” she said, and some use voices found in databases without paying for them.

“There’s that grey area” being exploited by several platforms, Taylan acknowledged.

“They take your voice, my voice, five other people’s voices combined that just creates a separate voice… They say that it doesn’t belong to anybody.”

Speechki, a Texas-based start-up, uses both its own recordings and voices from existing databanks, said CEO Dima Abramov.

But that is done only after a contract has been signed covering usage rights, he said.

Future of coexistence?

The five largest US publishing houses did not respond to requests for comment.

But professionals said several traditional publishers are already using so-called generative AI, which can create texts, images, videos and voices from existing content — without human intervention.

“Professional narration has always been, and will remain, core to the Audible listening experience,” said a spokesperson for that Amazon subsidiary, a giant in the American audiobook sector.

“However, as text-to-speech technology improves, we see a future in which human performances and text-to-speech generated content can coexist.” The giants of US technology, deeply involved in the explosively developing field of AI, are all pursuing the promising business of digitally narrated audio books.

Published in Dawn, May 15th, 2023