Less discussed than services like ChatGPT or Sora are the platforms built specifically to generate music. There are several on the market created for this very purpose — and their output has already begun clogging up streaming services. The streaming platform Deezer estimates that around 20% of new music uploaded to their site is AI-generated, and as these services become more popular, that number is only going to increase.

The saddest part? Unlike images or text, where people are slowly learning to distinguish human work from AI, virtually no one seems to be able to do it with music.

That’s not hyperbole. In a new study, Deezer asked 9,000 people across Brazil, Britain, Canada, France, Germany, Japan, the Netherlands and the United States to listen to three pieces of music — two generated by AI, one made by a human.

The result? A full 97% of people couldn’t tell the difference between music made by a human and music made by an artificial intelligence.

While this is disappointing, it shouldn’t come as that much of a surprise. There have already been several AI-music scandals this year alone, from AI band “The Velvet Sundown” garnering millions of listens to one of the top country songs in the nation being generated by AI.

Thankfully, people polled by Deezer said they didn’t like the fact that they couldn’t tell the difference between the two, showing that there’s at least a little distaste for AI content out there.

Of course, that doesn’t mean that these losers are going to stop creating it. If you thought that music was about expressing emotion, telling a story, or sharing a feeling that can’t be communicated with words alone — sorry, buddy, but now, everything’s computer!