Audiobook narrators say AI is already taking away business
AI can make records that sound like people, and it can do this quickly, bypassing at least some of the services of people who have made a living with their voices for years.
Many of them are already seeing their business drop sharply.
Tanya Eby has been a skilled voice actor and storyteller for 20 years. She records music in her own home.
But in the last six months, she’s had half as much work to do. In a normal year, her bookings would go through August. This year, they only go through June.
Many of her coworkers have seen similar drops.
Even though there could be other causes, she said, “It makes sense that AI is affecting us all.” There is no label that says something is an AI-assisted recording, but experts say that thousands of ebooks use “voices” made from a databank.
DeepZen has rates that can cut the cost of making an audiobook to one-fourth or less of what it would cost for a traditional job.
The small company in London uses a database it made by recording the sounds of actors who were asked to speak in different ways to show different emotions.
“Every voice we use is licensed, and we pay for the recordings,” said Kamis Taylan, the CEO of DeepZen.
He also said, “We pay royalties for every project based on the work that we do.” Eby said that not everyone follows that rule.
“There are all these new companies that aren’t as honest,” she said. “Some of them use voices from databases without paying for them.”
“There’s that gray area,” Taylan said, and several sites are taking advantage of it.
“They put your voice, mine, and the sounds of five other people together to make a new voice… They say that it belongs to no one.”
The CEO of Speechki, which is based in Texas, said that the company uses both its own recordings and voices from current databanks.
But, he said, that doesn’t happen until after a contract is signed that spells out the rights to use.
The future of getting along?
When asked for their thoughts, the five biggest US publishing companies did not answer.
But experts say that a number of traditional publishers are already using “generative AI,” which can make new texts, images, videos, and voices from existing material without any help from a person.
“Professional narration has always been and will continue to be a big part of the Audible listening experience,” said a spokeswoman for this Amazon subsidiary, which is a giant in the American audiobook market.
“However, as text-to-speech technology gets better, we see a future where human performances and content made by text-to-speech can live together.” All of the big tech companies in the US are trying to get into the business of digitally narrated audio books. These companies are all heavily involved in the fast-growing area of AI.