YouTube and the rise of artificial intelligence aren’t exactly strangers but we haven’t heard an awful lot that sounded like it wasn’t going to be an immediate dumpster fire (so far). That might change with two new AI tools for the video-sharing platform that could make life easier for creators. Or, you know, considerably more difficult.
These new features, according to a new blog post, are called Dream Track and Music AI Incubator. There’s more to these products than just the name, which tells you barely anything. The first allows creators to use artificially generated voices for 30-second clips. The second is all about music creation and it’ll apparently turn a hummed tune into an actual guitar riff (among other things).
Dream Track sounds like the sort of feature folks have been experimenting with ever since computers were taught to read aloud. The difference is that this time, YouTube creators won’t have to rely on a stilted artificial voice to narrate their Shorts submissions. A short list of somewhat notable artists are offering up their voices to help create 30-second clips of video narration.
The list of nine artists is Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Papoose, Sia, T-Pain, and Troye Sivan. Odds are you’ve heard of at least one of these. Well, some video creators will be able to press these voices into service to create the aforementioned clips. Using AI, of course. You don’t get to participate, at least not yet. The minds at YouTube know better than to unleash this sort of feature onto the world at large. Making someone famous perform the Castle Argh sketch from Monty Python isn’t the worst thing you could do but that’s just where it starts.
Music AI Incubator, on the other hand, is all about making music more effectively. And by ‘more effectively’ we really mean ‘as little effort as possible’. The ability to turn a hummed tune into a part of a song is just the headline feature, though YouTube also punts the ability to change a song you’re working on from one genre into another. We don’t know whether you’ll be able to re-render a catchy pop track into a deeply atmospheric funeral doom song but we’d certainly love to see the system give it a shot.
It’s got other features but YouTube isn’t expounding on them too hard just yet. Music AI Incubator will roll out to a first wave of users — likely as controlled as Dream Track is — in 2024. The moves make sense. It’s one thing to come up with a cool feature. It’s another to unleash tools that can further blur the line between truth and fiction with little regard for how they’re used. We’re certain folks will find a way to do wacky things when these features become widespread. It’ll depend on YouTube and just how much damage is possible when that day finally comes.