What if AI could read your audiobooks?
If you’re an audiobook fan, Apple just gave you a taste of the future by releasing its first batch of AI-narrated books. But while the initiative is fascinating and has far-reaching implications, the robotic tone of the narrators shows that the beloved human voice is here for the long haul.
You can find audiobooks read by an AI, using a text-to-speech system, in the Apple Books app by searching for “AI narration.” This will give you a list of romance or fiction books (free or paid) with the description “narrated by Apple Books.”
Apple Books offers two types of AI voices – a soprano called Madison and a baritone called Jackson – both of which have an American accent and currently only speak English. You can get a taste of the sound of them by pressing the “preview” button under any of the titles narrated by Apple Books.
Right now, Apple’s two AI voices have a distinctly robotic, artificial tone. You won’t soon confuse them with the warm and meaningful tone of famous storytellers like Stephen Fry or Julia Whelan. But while the voice pitch remains a difficult hurdle for AI storytellers, there’s no doubt they’re about to win our ears.
Naturally, Apple claims (opens in new tab) whose AI voices were developed to make audiobooks “more accessible to everyone”. But they’re also making the multi-billion dollar audiobook industry more accessible to the tech giant. And the new Apple Books feature is just the beginning of a fierce battle with Amazon and Spotify to win our audiobook-loving ears.
voice problems
For now, we’ll mostly see AI narrators limited to books from smaller independent publishers. This development could cause an explosion in the number of audiobooks available to readers on all devices, as digital storytelling opens up a new market for publishers and authors that, until now, was not able to switch from print to audio.
But the reluctance of major publishers and voice actors may also slow the rise of robo-narrators. Amazon’s Kindle e-readers officially lost their text-to-speech feature a few years ago, although there are workarounds in the accessibility menus. This decision is due, at least in part, to copyright issues and the fact that audiobooks are legally considered separate works of art.
Audible, owned by Amazon, also wrote at length (opens in new tab) about the types of narrators that suit different styles of books and how publishers choose the right ones. His blog states that “the most important thing about audiobooks is that the voice matches the tone and genre of the book.” It’s a bit difficult to achieve this goal if, like Apple Books, you only have two voices.
Audible also notes that “experienced voice actors capture listeners’ attention by adjusting pitch, intonation, volume and accent”. It is precisely in this area that AI readers should undergo intense vocal training, and perhaps also take evening classes in emphasizing emotions.
But the new Apple Books feature is clearly just the beginning of an inevitable boom in AI voice technology. And the big moment for audiobooks could be when AI can, instead of a robot reading a script, convincingly impersonate a famous voice actor – a breakthrough that may not be too far off, though. believed.
The fake empire
After all, “deepfakes” are not limited to the terrifyingly convincing videos of a Morgan Freeman (opens in new tab) or a Tom Cruise (opens in new tab) synthesis: AI-powered voice technologies are also developing rapidly.
Viral web app like Uberduck (opens in new tab)allows you to compose speeches with the voices of former presidents or cartoon characters, while last year Amazon introduced (opens in new tab) a slightly scary new Alexa skill that can read the Wizard of Oz to a child in their grandmother’s voice.
With companies like Google Wavenet (opens in new tab) that also advances in technology, AI voices can only become more compelling. For now, the barriers to widespread use of AI in audiobooks are likely to be legal and ethical, rather than technological. But Apple Books’ artificial narrators are the sound of our synthetic future — and in the not-too-distant future, popular voice narrators will likely be copyrighted and licensed for their own voices as well.
So far, data from Apple Books suggests that these AI voices are better suited to nonfiction and factual works than moving narratives. For novels, as for films, we are still far from seeing artificial actors convincingly strike our chords without ruining the suspense with a robotic inflection or a flat note. But a torrent of AI-read audiobooks will come, no matter what.
