The ways in which audiobooks are being made and consumed are in a state of dizzying flux. Artificial intelligence (AI) narration is rapidly becoming a viable alternative.
In fact, it already is a viable alternative: especially for nonfiction, but for fiction, too. (Fiction will just require a little more post-production editing and adjustment.)
Don’t take my word for it. At least two major retailers (Apple and Google) agree. Both companies have rolled out AI narration for titles listed in their online stores as ebooks. Both Apple and Google have invested in in-house technologies for this specific purpose. This means that they see AI narration as the future.
With these olives disgorged from the bottle, big changes in the audiobook landscape are practically inevitable.
As is always the case with change, there will be tradeoffs. Not everyone will be happy with the outcomes.
Let’s start on the production side. For creators, the shift to AI narration might be either exciting or disastrous, depending on which side of the supply chain one finds oneself. Some people will lose their current livelihoods. Others will earn a lot more money as a result of lower production costs.
The future may be difficult for audiobook narrators who aren’t named Scott Brick or January LaVoy. On the other hand, indie authors and small publishers may soon be able to profitably produce audiobooks for low-volume, “long tail” titles.
Consumers will notice changes, too. For audiobook listeners, the changes will mostly be positive. Some consumers will be thrilled with all the new choices available. Within a few more years, there will be thousands more audiobooks on the market, and they should be a lot cheaper than they currently are.
But not all of these new AI projects will be delivered with the same passion and emotional clarity of the best human-narrated audiobooks. This will be less a result of technological limitations, than the inevitable gold rush mentality that may soon grip the audiobook market.
It is now possible to create an AI audiobook that doesn’t sound like the deadpan voice in a 1980s video game. (I’ve used the Google AI narration tool, so I’m speaking from a position of knowledge here.)
But making technology sound human is a process that takes time and effort. It always will. Artificial intelligence isn’t truly “intelligent” like a human being is intelligent. If you want a decent AI recording, you have to listen to each word in the editing phase, and make decisions about its pronunciation and cadence. There is no way to accomplish that by simply pushing a button, or by typing in a few commands.
As is always the case with things that require effort, there will be folks who try to skip essential steps, and rush shoddy products to market. The “tsunami of crap” alarm bells are already ringing. And in some cases, at least, the bell-ringers will have a point.
These your-mileage-may-vary tradeoffs are unavoidable. So is the change. I’m not saying that all this change is a good thing or a bad thing. But if you have any interest in audiobooks—as a writer, narrator, or consumer—you would be wise to learn about these changes. Because they’re definitely coming.
