Video: Rick Beato
For someone as profoundly unmusical as me, AI music generators are quite magical. I can barely sing a note, but in a few seconds I can make an entire track in any genre on any topic I want – like this soul song about Sky camera operator Phil Hooper. You can dismiss this as pure silliness for an audience of about five, but to me that’s the point! Thanks to AI, I get a little bit of musical joy that otherwise is completely out of reach.
Yet, as ever with technology, removing friction comes with a cost, and in this case the cost is a tsunami of musical spam. The stats on AI music are mind-boggling. In 2015, the entire US music industry made around 57,000 songs. Today, 60,000 AI tracks are uploaded to Deezer (aka French Spotify) *every single day* - that’s 21m a year, and this thing is just getting going.
The real problem isn’t the tracks, however, but the behaviour around them, because AI music is being used to try and steal from streamers (and by extension every legitimate musician on the site). Deezer estimate that 85% of listens to AI music are fraudulent – that is, made by bots set to stream the songs over and over in order to siphon royalties from the common pool.
Yet, as ever with technology, removing friction comes with a cost, and in this case the cost is a tsunami of musical spam. The stats on AI music are mind-boggling. In 2015, the entire US music industry made around 57,000 songs. Today, 60,000 AI tracks are uploaded to Deezer (aka French Spotify) *every single day* - that’s 21m a year, and this thing is just getting going.
The real problem isn’t the tracks, however, but the behaviour around them, because AI music is being used to try and steal from streamers (and by extension every legitimate musician on the site). Deezer estimate that 85% of listens to AI music are fraudulent – that is, made by bots set to stream the songs over and over in order to siphon royalties from the common pool.
[ed. Do check out the soul song example mentioned above (with this accompanying video). Pretty scary... and sad.]
