So YouTube’s plan for stopping AI-fueled plagiarism is “we’re working on it”?

Everyone had heard of ChatGPT here and there, but most people only saw it as Cleverbot 2.0. I mean, how much more impressive can the thing that powers those funny talking saltshakers from YouTube be?

Well, as it turns out a lot! Especially the part related to generative AI, where the platform can create something new, yet not entirely original. Like, applying the art style of a given artist to make a portrait based on several of your selfies.

And as it turns out, you all liked it so much that companies like Microsoft and Google started putting heavy emphasis on the technology. So much so that we’re expecting AI-fueled features to pop-up on some of the best phones to come out soon.

But there’s a dark side to this too. Because, doesn’t the artist that you based the art style for your AI-generated portrait deserve some credit? What about singers whose styles AI copies for half-original songs that found themselves on YouTube or Spotify?

Image credit – PhoneArena

Well, given that a huge chunk of Google’s I/O event of 2023 was practically dedicated to AI — besides the reveal of the Pixel 7a and the Pixel Fold — is it any wonder that YouTube — which, as a kind reminder, is also owned by the Big G — laid out its plans to combat AI-fueled plagiarism in this blog post?

And before we move on to said plans, kudos to CEO Neal Mohan for acknowledging AI’s application as a creative tool too!

Okay, so what’s the plan to combat bad AI actors? Well, it’s called YouTube’s Music AI Incubator. The name itself isn’t really descriptive of what it should be doing. Truth be told, details such as the included technology or required process have been left out for “some time in the future” entirely.

What we do know is that YouTube plans to partner with huge musicians from all around the world — starting off with Universal Music Group, a giant that is quite opposed to user-generated songs — in order to conduct experiments related to generative AI… To some sort of end related to the aforementioned incubator. Cool?

There is also mention of planned work, which would expand the way that Content ID works on YouTube — this is the thing that gets you striked for using copyrighted music, by the way. Basically, using AI, the system will be able to distinguish between the original artist and their inclusion in user-generated content, ensuring — in theory — proper distribution of monetization.

And if that sounds difficult, it’s only because it is. But look at it this way: if it all works out and you make an AI song with Drake’s voice, this would mean that Drake can get his share, while you can also get something on the side, if the song blows up.

And then there is a short mention about YouTube’s planned defense against fake news, because — as you’ve probably already seen — AI can also make pretty convincing deep fakes of social and political figures.

The plan there? “We’re working on it”, but explained in about three paragraphs. Neat!

It’s always great to see that a huge platform like YouTube, where a lot of artists from all sorts of calibers are making their living, is concerned about the future. That being said, to me, this post felt a bit “undercooked” and I’m certainly looking forward to checking out the specifics of the plan, when it is in motion.