Is that song real? YouTube's new tech will catch it if it's not

Ever wondered if the track you’re listening to on YouTube is actually real? Just a few years back, it seemed impossible that an AI-generated song could sound so convincingly human, but that’s exactly where we find ourselves now. While it’s getting pretty hard to tell the difference between what’s real and what’s AI, YouTube has a few ideas on how to tackle the problem.

YouTube is building new technology within Content ID (its system for identifying copyrighted material) that claims to be able to spot synthetic singing. If you’re a creator and someone tries to clone your voice with AI and use it in a song, YouTube’s system is designed to catch it.

Alternatively, if you’re a casual YouTube viewer, this tool could help reassure you that the music you’re enjoying on the platform is authentic and not some AI-generated knockoff that’s piggybacking off someone else’s identity and hard work. YouTube is currently working with different partners to fine-tune this tool and states that a pilot program will be available early next year. If all goes well, we should see this technology rolled out widely soon after.

The company also shares that it’s developing tech that will help everyone from creators, to athletes and actors, spot AI-generated content that deepfakes their faces without consent. While we’re not sure how AI will evolve in the future, these features that YouTube is developing feel like the right sort of tools we want to have alongside it. This isn’t just a tech upgrade for the sake of it or a feature rollout to keep up with competitors; rather, it’s a solid way to protect the likeness of creators and, in turn, ensure that users have access to genuine content on the platform.

With privacy concerns around AI growing, it’s this sort of tech that has the potential to make a real difference. We’ll have to wait and see how well YouTube’s solution works and whether it can truly help keep deceptive AI content off the platform.