As AI fakes proliferate, we need to draw a clear distinction between human-made music and AI-generated contentEver wondered what’s in your morning cereal? Easy – just turn the packaging over and, by law, you’ll see the maker’s clear list of ingredients. Flick on a morning playlist, however, and there’s no immediate way to tell if the song you’re listening to is real or AI-generated.
AI content is filling up streaming platforms. Some mimic the voices of living artists, some the voices of the deceased, and some are batch-generated and posted under the name of entirely fictional artists. All of it draws revenue away from working musicians.
A recent study by MusicMagpie, estimates that, on YouTube alone, there are already well over 1.5 million AI cover tracks. After analysing the 1,500 top-performing AI tracks, they found that Blackpink, Justin Bieber, and Kanye West were out in front as the most AI-plagiarised artists, racking up over 33 million views between them.

READ MORE: How will young people learn music and production in a post-AI music industry?

The study estimates that the music industry has lost over £10 million in revenue to AI-generated content, but this is likely just the tip of the iceberg. Consider that one North Carolina musician, with a relatively small-scale operation, was allegedly able to generate hundreds of thousands of fake tracks, rack up billions of plays, and syphon off millions of dollars from streaming platforms. The true scale of revenue loss already occurring is likely far higher than we know.
Artists and labels are understandably upset by all this, with many voicing their support for the No AI Fraud Act in the US, and thousands more signing an open letter to protest the dodgy business practices that make much of this possible. Crucially, music fans are unhappy about the situation as well.
AuthorityHacker surveyed 1,200 music consumers and got some striking results: 93% said they did not value AI-generated music as highly as music produced by humans. And, while over 60% said they would consider listening to AI music, some 56% also said they would not willingly pay for songs generated using AI. Perhaps most striking: 89% of those surveyed want clear labelling for AI-produced music.
Image: Kemal Yildirim via Getty Images
Are these numbers surprising? Not really. Think of the perceptual difference between champagne and sparkling wine; can most drinkers actually taste the difference between a high-priced bottle of bubbly and a cheap one? No. But people want to know what region their wine came from and are perfectly willing to pay more for the real deal. That stamp of authenticity is enforced by government action that ensures winemakers are protected from counterfeiters. So, where is the protection for music makers?
We will soon arrive at a place where AI-generated music can match the catchy-ness, complexity, and audio fidelity of what human creators can do. Before we reach this parity, governments and industry must work together to put clear labelling rules in place and empower music consumers to make an informed choice regarding what kind of music they want to patronise.
Mandatory labelling for vocal deepfakes and wholly generated content is the obvious first choice, but an argument can be made that we shouldn’t stop there. Record labels, and even some artists, might baulk at the prospect of having to disclose whether a song was produced using AI mixing and mastering tools, but professionals at all levels of the industry – not just those at the top – deserve some protection from AI counterfeiting their work.
Music is not a luxury, it is woven into our society. Take the UK for example: for generations, its music scene has been hugely influential and a major commercial force, with UK artists routinely among the top-selling worldwide. To let that legacy drown in a sea of AI sludge would amount to cultural manslaughter.
Sure, the idea of top-shelf, organic, sustainably sourced music may seem slightly ridiculous – but a brief glance at the booming vinyl market provides plenty of evidence that people will pay more for something they believe has real value. Given a clear choice between human and machine-made music, there is every reason to suspect that the listening public will go for the flesh-and-blood option.
The post As AI fakes proliferate, we need to draw a clear distinction between human-made music and AI-generated content appeared first on MusicTech.