When we think about what makes an audiobook memorable, it’s always the most human moments: a catch in the throat when tears are near, or words spoken through a real smile.

A Melbourne actor and audiobook narrator, Annabelle Tudor, says it’s the instinct we have as storytellers that makes narration such a primal, and precious, skill. “The voice betrays how we’re feeling really easily,” she says.

But as an art form it may be under threat.

In May the Amazon-owned audiobook provider Audible announced it would allow authors and publishers to choose from more than 100 voices created by artificial intelligence to narrate audiobooks in English, Spanish, French and Italian, with AI translation of audiobooks expected to be available later in the year – news that was met with criticism and curiosity across the publishing industry.

In Australia, where there are fewer audiobook companies and where emerging actors like Tudor rely on the work to supplement their incomes, there is growing concern about job losses, transparency and quality.

While Tudor, who has narrated 48 books, isn’t convinced that AI can do what she does just yet, she is worried that the poor quality may turn people away from the medium.

“I’ve narrated really raunchy sex scenes – AI doesn’t know what an orgasm sounds like,” she says. “Birth scenes as well – I’d love to know how they plan on getting around that.”

The audiobook boom

According to a 2024 report by NielsenIQ Bookdata, more than half of Australian audiobook consumers increased their listening over the past five years. Internationally there was a 13% increase in US audiobook sales between 2023 and 2024; in the UK audiobook revenue shot up to a new high of £268m, a 31% increase on 2023, the Publishers Association said.

As demand for audio content grows, companies are looking for faster – and cheaper – ways to make it. In January 2023 Apple launched a new audiobook catalogue of audiobooks narrated by AI. Later that year Amazon announced that self-published, US-based authors with works on Kindle could turn their ebooks into audiobooks using AI “virtual voice” technology – and there are now tens of thousands of these computer-generated audiobooks available through Audible.

And in February this year, as part of a more general shift towards audiobooks, Spotify said it would be accepting AI audiobooks to “lower the barrier to entry” for authors hoping to find more readers.

Audible says its aims are similar: to complement, not replace, human narration, allowing more authors and more titles to reach bigger audiences. In the US Audible is also testing a voice replica for audiobook narrators, to create dupes of their own voices that will “empower participants to expand their production capabilities for high-quality audiobooks”.

“In 2023 and 2024, Audible Studios hired more [human] narrators than ever before,” an Audible spokesperson told the Guardian. “We continue to hear from creators who want to make their work available in audio, reaching new audiences across languages.”

But robot narrators will always be cheaper than humans – and people in the voice acting and book industries fear a move to AI could pose a threat to workers.

Volume or quality?

Dorje Swallow’s career as a narrator took off after he began voicing novels by the Australian bestselling crime author Chris Hammer – and the actor has now narrated about 70 audiobooks. Swallow believes AI narration is a tool created by people who “don’t understand the value, technique and skills” required to produce quality audiobooks.

“We’ve done the hard yards and then some to get where we are, and to think you can just press a button and you’re going to get something of similar, or good enough quality, is kind of laughable,” he says.

Simon Kennedy, the president of the Australian Association of Voice Actors, says there has always been a battle over how much a narrator deserves to be paid in Australia. For every finished hour of an audiobook, a narrator might spend double or triple that time recording it – and that doesn’t include an initial read to understand the book and its characters.

“My personal opinion is that [introducing AI narrators] is going for volume over quality – and it’s looking to cheapen the process,” he says.

Kennedy founded the Australian Association of Voice Actors in 2024 in response to the threat being posed by AI. In a submission to a parliamentary committee last year the organisation said 5,000 Australian voice acting jobs were at risk.

He was hardly surprised about Audible’s announcement but says he thinks it’s a “pretty dumb move”.

“An audiobook narrator has such a special and intimate relationship with the listener that to try and do anything that is less connective is a foolish move,” he says.

As for the opportunity to clone their own voices, he says voice actors should have the right to engage – but they shouldn’t expect “any near the same pay rate, and they risk turning their unique timbre – their vocal brand – into a mass-produced robot voice that listeners get sick of listening to pretty quickly.”.

“If an emotionless narration at a consistent volume is all you need for ‘high-quality’, then sure,” he says. “But if engaging, gripping, edge-of-your-seat storytelling is your version of high-quality, then don’t hold your breath for AI to give you that.”

Another major concern is Australia’s lack of AI regulation. While the EU has its own AI Act, and China and Spain have labelling laws for AI-generated content, Australia is falling behind.

“There are no laws to prevent data scraping or non-consenting cloning of voices, or of creating deepfakes of people,” Kennedy says. “There are also no labelling laws or laws to mandate watermarking of AI-generated content and its origins; no laws to mandate transparency of training data; and no laws to dictate the appropriate use of AI-generated deepfakes, voice clones or text.”

This year the Burial Rites and Devotion author, Hannah Kent, was one of many acclaimed Australian writers shocked to discover their pirated work had been used to train Meta’s AI systems. She says while her initial reaction to the introduction of AI into creative spaces tends to be “refusal and outrage”, she’s curious about Audible’s AI announcement – specifically its plans to roll out beta testing for AI to translate text into different languages.

“I think it’s fairly obvious that the main reason to use AI would be for costs, and I think that’s going to cheapen things in a literal sense and cheapen things in a creative sense – in that sense of us honouring the storytelling, artistic and creative impulse,” Kent says.

Tudor and Swallow believe big companies will struggle to replace human narration completely, partly because many Australian authors will oppose it.

But whether or not listeners will be able to tell the difference remains to be seen.

“The foot is on the pedal to drive straight into dystopia,” Tudor says. “Can we just listen to people instead of robots?”