AI Voiceovers – are they any good?
There’s no escaping it, AI-generated voiceover tracks are now unavoidable. They’ve improved hugely, and quickly become an invaluable tool in the video production process. But are they actually any good? The answer is a resounding yes, but also kind of…. No.
Why they’re good
AI voiceovers are quick and easy to create. Using either a text prompt or an existing audio reference track, you can have a pro-sounding voiceover ready in seconds.
This means they’re cheaper than working with a human VO artist, and a great choice for quick turnaround projects, those that are likely to have lots of script revisions, or when budget is particularly tight.
They’re useful as a placeholder for testing out scripts ahead of a proper VO recording, or for boosting accessibility on videos where you might not normally include a VO.
Why they’re bad
Robots are getting better at conveying emotion. But while you do have some control over delivery and pronunciation with AI voiceovers, it’s limited, and the more granular you try to be with this, the more you lose the benefits of speed and cost. The fact is if nuance and expression is critical for your script, you need a real human to provide it.
There’s also the fact that people’s ears are already becoming attuned to AI voiceovers, and they’re increasingly able to tell when they’re listening to a robot. In certain contexts this can distract from a video’s messaging, or make it feel impersonal.
Our view
We’re finding AI-generated voiceovers to be a hugely useful tool in the video creation process, especially in the early stages. There also open up exciting possibilities, like automatic translation to different languages, or using voice imprints to allow a person to sound like someone entirely different.
But for now, we see them as something that won’t replace traditional voiceover artists in the production process, but sit alongside them; robots and VO artists, together in narrative harmony.
