AI Audiobook Narration: How Will It Impact Voice Artists?

AI audiobook narration isn’t just coming, it’s already here. With The Guardian reporting that Apple has quietly introduced its first suite of AI-voiced books, and Audible set to do the same, this is a pressing concern for all voice artists. As such, we wanted to discuss the current situation and how AI audiobook narration might change the industry.

AI is an unmistakable threat to the arts. Where previously AI has been something of a scientific novelty, it is now reaching a point where it can convincingly imitate human work. AI art can be used instead of employing artists, AI writing can be generated, and AI voices can be used to auto-narrate.

But can an AI voice ever match a human? And if so, could they ever become the norm?

AI Audiobook Narration: Can an AI Voice Match a Human?

The short answer to this question is ‘yes’. The tech used for AI projects is advancing by the day. Whilst the tech may not specifically be designed for AI audiobook narration, once it exists it will be in the interest of companies to use it.

The days of the stilted ‘computer voice’ we associate with generated voices are already long gone. Just as ‘deepfakes’ (software that can edit a video to make a person look like someone else) are already a concern, we will soon reach the point where AI voices are indistinguishable from real ones.

That doesn’t mean that voice artists will immediately have their work snatched away by AI software. The work that goes into an audiobook can’t be auto-generated.

Most narrators will read the book several times, first for context, and then to specifically mark up the text. They need to create distinct character voices based on descriptions and dialogue, and check pronunciations of specific words. That is usually several days of work, and it’s not a task that can be automated.

Once the recording starts, there are still advantages to human readers. Human readers bring emotion to line readings and can set mood with their voice. Even on a technical level, a human knows when they’ve made an error and can correct it. Any generated reads would require someone to make those checks and corrections. And if an AI voice cannot pronounce a word as it is meant to be pronounced (or in the way the text requires) this is a more significant tech problem than having a human immediately try again.

AI Audiobook Narration: Should I Be Concerned About Using my Voice for AI Projects?

AI’s impact on the voice industry isn’t just confined to audiobooks. As voice technology improves, AI voices are being used for advertising and corporate videos, with companies offering suites of voices that can narrate any typed text for a fee.

This will have a huge impact on the industry, as it’s likely that many clients who don’t need human voices (for presentations, social media posts) may eventually move to these services.

This isn’t necessarily reason to panic. Many of the clients using those services are likely to have smaller projects with budgets that wouldn’t have allowed for professional narration to begin with. The danger for the industry is when larger clients who currently prefer voice artists make the switch to voice banks.

Most voice artists have seen a rise in AI related jobs being advertised, and many of them are for this purpose, though it’s not always obvious. The fees can be large but require a huge amount of recording time (nearly always recorded and edited in a home studio). The money may be tempting, but the contracts are often vague. While they may couch usage in terms of ‘research’ or ‘development’ it may, in reality, mean signing away the rights to your voice to anyone who pays the company to use it.

There are plenty of AI voice projects that have more limited use (many with perfectly above-board scientific or tech-uses!) We suggest always checking the contract and being one-hundred per cent clear as to the purpose and use of the recording. The decision of how you want your voice to be used is ultimately up to you.

AI Audiobook Narration: Should Voice Actors Worry?

The recent news about Apple and Audible’s introduction of AI audiobooks was met with protest. In the same way Science Fiction publisher Tor was met with universal disapproval when they used AI generated art for a book cover, instead of employing an artist. AI is a huge threat to the voice industry and it is important that artists, agents and unions stand against the widespread use of AI narration.

In its current state, the use of AI audiobook narration being introduced by Apple and Audible is more likely to be sold as a service to the huge percentage of self-published authors that are published by both companies. These would be smaller-budget or self-published books which would never ordinarily have had an audiobook version under other circumstances. In its current state it cannot be a high-quality option for publishers and will not be a popular choice for listeners.

AI seems an unstoppable force across many areas of life, but just because the tech exists, it doesn’t mean it’s going to be in demand. There is the famous ‘Uncanny Valley’ effect in which CGI-animated faces can cause feelings of repulsion in humans. The mental health app Koko introduced AI chatbots and quickly withdrew them; when a chatbot says ‘I understand’ or ‘I’m sorry to hear that’ it doesn’t have the same emotional impact.

Before there were written stories, there were storytellers. The human need to be told a story has never gone away, and it’s very hard to imagine that it ever will. We may very well end up getting used to the ‘new normal’ of AI voices, but audiobook narration will still be dominated by humans.

A generated voice can’t falsify the emotion of a love story or invent the magical world of a children’s book. No matter the tech, the human voice isn’t going anywhere.

At OutSpoken we can connect you with a wide range of voices from around the world. Each have a high-quality home set-up and are professional voice artists with proven industry experience. If you are seeking an affordable, versatile voice actor for an audiobook, please do check out our list of voice talent or get in touch.

January 18th at 12:00am