The reasons audiobooks resonate are deeply human. They evoke moments that catch in the throat or a genuine smile when a word is spoken.

Melbourne actor and audiobook narrator Annabelle Tudor refers to narration as a storyteller’s innate ability, a fundamental and priceless skill. “The voice easily reveals our true feelings,” she explains.

However, this art form might be facing challenges.

In May, Audible, the audiobook service under Amazon, revealed plans to enable authors and publishers to narrate in English, Spanish, French, and Italian, using over 100 voices generated by artificial intelligence.

With a dwindling number of audiobook companies, emerging talents like Tudor are increasingly reliant on these opportunities, sparking concerns regarding job security, transparency, and overall quality.

Having narrated 48 books, Tudor is uncertain whether AI can replicate her work, yet fears that a dip in quality may alienate listeners.

“I once narrated a particularly explicit scene. The AI lacks understanding of how an orgasm sounds,” she remarks. “I’m curious to know how they plan to address such nuances, including more delicate scenes like childbirth.”

The Audiobook Boom

A 2024 report from Nielseniq Bookdata indicates that over half of Australian audiobook consumers have increased their listening in the last five years. On an international scale, US audiobook sales have risen by 13% from 2023 to 2024. Meanwhile, the UK has seen audiobook revenues soar to £268 million, marking a 31% increase in 2023, as reported by the Publishers Association.

As demand for audio content surges, companies are seeking quicker and cheaper production methods. In January 2023, Apple unveiled a new catalog featuring AI-narrated audiobooks. Later that year, Amazon introduced a feature allowing self-published authors to convert their Kindle ebooks into audiobooks using “virtual audio” technology, resulting in tens of thousands of Audible titles now available in AI-generated formats.

Additionally, in February, Spotify announced support for AI audiobooks, making it easier for authors to reach wider audiences. Audible claims its goal is not to supersede human narrators but to enable more authors and titles to connect with larger audiences. In the US, Audible is testing audio replicas of audiobook narrators to create a unique voice, enhancing their capacity to produce high-quality audiobooks.

“In 2023 and 2024, Audible Studios has hired more [human narrators],” a spokesperson shared with the Guardian. “We continually engage with creators eager to have their work available in audio format and reach new audiences across languages.”

Yet, robot narrators remain a more economical choice than human talent, raising fears among industry professionals about potential job threats.

Volume vs. Quality?

Australian bestselling crime author Chris Hammer helped narrator Doge Swallow launch his career, highlighting a belief that AI narration is a tool designed by people who fail to grasp the intricacies, techniques, and skills necessary for quality audiobook production.

“Some assume we just press a button for a similar or sufficient quality result,” he notes.

Simon Kennedy, president of the Australian Audio Actors Association, mentions a long-standing struggle in Australia about fair remuneration for narrators. Recording an audiobook can mean narrators spend up to three times the length of the finished product for recording, not counting the initial read to understand the narrative and characters.

“In my view, AI narrators prioritize volume over quality and aim to cut costs,” he asserts.

In 2024, Kennedy founded the Australian Voice Subject Association in response to AI’s looming threat. In a submission to a parliamentary committee last year, the organization warned that 5,000 Australian voice acting jobs were at stake.

While not surprised by Audible’s recent announcement, he dismisses it as a “foolish decision.”

“Audiobook narrators hold a truly special and intimate connection with their listeners; pursuing an approach that lacks this connection is misguided,” he suggests.

Regarding voice cloning opportunities, he states that voice actors should be involved in the process, but warns that it may lead to a homogenized robotic voice that listeners quickly tire of.

“If a monotonous, emotionless narration suffices for ‘high quality,’ then perhaps,” he counters. “However, if you seek an evocative, captivating listening experience, don’t expect to find it there.”

Another pressing concern is the absence of AI regulations within Australia. The EU has its own AI ACT, while China and Spain also have measures in place, whereas Australia lacks regulations regarding the labeling of AI-produced content.

“No laws exist to prevent data scraping, voice cloning, or breeding deeper AI capabilities,” Kennedy explains. “There’s no labeling or transparency requirement for AI-generated material or its origins, nor any regulations governing the proper use of AI-generated deepfakes, audio clones, or text.”

This year, during the burial ceremony and dedication of her work, Author: Hannah Kent dropped with astonishment upon discovering that pirated copies of her work had trained meta AI systems. Despite initial resistance and frustration towards AI’s infiltration in creative spaces, she shows curiosity about Audible’s AI developments and the prospective trials for translating texts into various languages.

“It’s evident that the primary motive behind AI adoption is cost-efficiency. Its aim is to reduce artistic value and creative narratives,” Kent reflects.

Both Tudor and Swallow agree that large corporations struggle to fully substitute human narration, as many Australian authors express opposition.

Yet, it remains unclear whether audiences can discern the difference.

“We are rushing straight into a dystopia,” Tudor warns. “Will I listen to humans or robots?”