One of the best things about the internet is having a seemingly endless supply of information available to you. News, books, films, shows, and music; you name it, the internet has it. You might have to pay for it, but if you wanted to explore the world from behind a device, the internet has you covered, especially if it was created by a human.
In this day and age, however, the human element is changing. We use computers to make stuff, and depending on how much of the process is handled by computers, we might even be entrusting computers to make it entirely without humans playing much of a part.
That’s where AI music comes into things, as models trained on human-made music can then be used to script sound in such a way where it can be difficult to discern the difference between real and AI-made. It can be so awkward that depending on how good the prompting and subsequent mixing is, you might find the real is almost indistinguishable from music made by AI.
While AI music isn’t used by everyone, it is properly flooding where we get our music from. Not so much the airwaves — radio is still largely controlled by people — but more the streaming services, which are seeing incredible increases.
This time last year, Deezer noted that 18 percent of all uploaded content was AI-made, up from 10 percent back in January. One year on, that number has more than doubled, exploding to a staggering 44 percent of music uploads.
To put that into real-world numbers, the music platform receives around 75,000 AI-generated music tracks each day, translating into around two million each and every month.
Or to put it simply: nearly half of the music uploaded to services now is AI-generated, and likely to turn a profit.
Some artists may get rich off the quality of their releases, going gold and platinum based on the writing, instrumentation, production, and engineering, but the people behind AI-generated music may be uploading in bulk to make money differently. The more plays they get, the more cents they can turn around, and eventually that leads to dollars and an income of sorts.
In terms of combatting the AI increase, music services including Deezer appear to be turning to tagging to indicate when something is made with AI. It’s something Apple Music supports with meta-data with “Transparency Tags” and a clear indication as to whether you’re listening to (and potentially supporting) music from an artist, or music from a service and system.
Depending on your point of view, that may lead you to switch off AI music, or at least bulk generated AI music, because there are clearly differences between the two. For instance, while there is a heap of AI music, commercial AI releases such as the most recent Beatles track may not note that it is made with AI, because it would have to be — only two band members are still alive, and the song was engineered using AI because of that.
It means that not every AI track will clearly be labelled because of this, with larger studios likely forgoing the transparency, but accounts uploading hundreds to thousands of tracks likely playing by the rules, or being forced to do so.
For listeners, the result remains largely the same: find the tracks and albums and genres they want to, and hopefully enjoy the outcome. Hopefully it won’t see you swamped by generative clones, though given the amount of uploads we’ve seen in the space of a year, we expect only more AI music to come.
