Who Are We Kidding
Music always seems to go first in tech/media disruptions. It was our files that blazed the way for digital sharing and piracy, paid downloads, unpaid streaming, and now fully generative AI. Last week, the streaming platform Deezer and data research firm Ipsos announced results of a survey about AI music:
“All participants were asked to listen to three tracks and determine whether or not they were fully AI-generated – 97% of the respondents failed.”
This isn’t the future of music. This is the now. We’re already in the Napster moment, the iPod moment, the Spotify moment with regard to AI music. It’s all around, whether you realize it or not. (97% may not be able to tell the difference, after all.) Last week Deezer also announced that 50,000 fully generative AI tracks are uploaded to the platform each day, accounting for more than a third of its new music. As Stuart Dredge of MusicAlly pointed out, that number is up from 10,000 tracks a day in January 2025; 20,000 in April; and 30,000 just in September. The trend hardly needs analysis. We’re already drowning in slop.
Which means it’s too late for most of the conversations I hear around me about music and AI. The question isn’t what are we going to do about it. I believe the more productive conversation at this point is what are we going to do afterwards.
Perhaps that’s my native optimism showing, or just diehard dialectics: AI will collapse from its own contradictions, whatever we do or don’t do about it. That doesn’t mean AI will disappear – on the contrary, it is here to stay like all technologies. The story of Pandora’s box isn’t about closing it. AI music is already a part of the technological landscape we live in; however, it won’t dominate that landscape forever. In fact, I would suggest its dominance may be very brief indeed.
Why? Cause it sucks.
I’m not going to link to the fully generative AI “country” song that supposedly topped a Billboard chart recently, cause I don’t want to give it any more clicks. Let’s just assume it’s as bad as you imagine, cause in fact it’s worse. As is most everything that AI churns out. It’s beyond imagination because it doesn’t come from one.
And yes, I suppose it is hypothetically possible that the corporations developing AI decide to steer the tool instead toward truth and beauty. Just as it’s hypothetically possible that Spotify decides to fulfill its stated mission, “To unlock the potential of human creativity by giving a million creators the opportunity to live off their art.” (Still a part of their staff recruiting, believe it or not!) But come on. It’s not happening. More importantly, it already didn’t happen.
When I look past AI, I feel rather cheerful. Because what AI is bad at, necessarily, is what humans are good at. Could there be a more positive development, at this absurd moment of capital spinning out of control, than a rededication to the value of human labor?
Music may again – I hope it will again – go first in this regard. Because what puts music in the forefront of so many technological trends is the same thing that rescues it after each debacle: music is simple to make, easy to share, and universally enjoyed. Fully generative AI music will quickly fill up all the space there is for audio slop. And we can take back the rest. All we have to do is survive this moment.
Stop waiting for it to happen, cause it already did. Start rebuilding. Now.
Listening to: Ellen Fullman, Elemental View
Cooking: Dried beans gleaned from dead plants
Jacques Attali’s book ‘Noise: The Political Economy of Music’ is very good on the subject of why music always goes first in disruptions.
It evokes that psychologist Harry Harlow who replaced monkeys with wire-shaped mothers, the bottle-holding surrogates, and compared the lives of their babies with those raised by the wire surrogates covered in cloth and others raised by living mothers.
I don't want to be sung to by a computer program that's slipped into the playlist without labeling or notification.
"If music be the food of love, play on." - William Shakespeare