Can you trust your ears to tell man from machine when it comes to music? Most people can't, it seems.

In a recent study by the streaming platform Deezer and market research company Ipsos, 97% of respondents could not tell the difference between musical tracks made entirely by artificial intelligence (AI) and those made by humans. In other words, AI-generated audio is really convincing. 

And it's already everywhere, including on the billboard charts: "Walk my Walk," from Breaking Rust — an entirely AI creation, from instrumentals to vocals to image — reached number one on the country digital song chart in mid-November. Then there is Xania Monet, a similar virtual artist with Gospel and R&B charting singles who was recently signed for $3 million (€2.6 million) to a record label. And this summer, the band Velvet Sundown gathered 1 million Spotify monthly listeners before revealing itself as a "synthetic music project."

Do people react negatively to AI music? 

At first glance, it might seem that listeners are welcoming AI-generated music with open arms and ears. But the truth is muddier. The same Deezer study found that 52% of respondents were uncomfortable about not being able to tell the difference between human and AI music. 

So, how do listeners feel when they know a piece of music is AI-generated? Some studies have found that they like certain types of music less. Other studies, including one by Philippe Pasquier, the director of the Metacreation Lab for Creative AI at Simon Fraser University in Vancouver, have found no bias for or against AI-generated music.

Part of this is because generative AI can be used in a myriad of ways that are not always clear to the listener. "It depends on what AI is used for — composition, interpretation, mastering — among other factors," Pasquier explained in a written interview with DW.

The ethics of listening to music

Sophia Omarji says she can enjoy an AI-generated piece of music. A Stockholm-based AI user experience researcher and music psychologist who hosts The Sound Mind podcast, she is also a musician. She told DW that while she thinks knowing a piece is AI-generated can make you want to "pull it apart," it doesn't technically change that much.

"You still enjoy the piece of music, but there is the kind of ethical and moral thing that might come [and make you think]: 'Is this something that I want to listen to?'"

One commonly raised ethical issue is that generative AI music platforms such as Suno and Udio train their machines on existing human artists' work. This potentially violates copyright laws for usage without payment. 

This has moved many artists to protest, including former Beatle Paul McCartney, who in December will release the single "Bonus Track." The recording of an empty studio will be part of the silent album, "Is This What We Want?" Created by over 1,000 co-writers that also include Kate Bush, Damon Albarn and Tori Amos, the release protests UK AI copyright legislation that artists say will devastate the music industry.

Can machines mimic human expression? 

Human fear of technological disruption is nothing new. As early as the 16th century, people were fearing job displacement as automation was introduced into knitting.

Forays into music-producing AI software have been around — and criticized —since the early 1980s. In recent years, AI has been used to finish Beethoven symphonies and Beatles tunes alike.

But to many musicians and music lovers, the current moment feels different.

"It's not just the loss of work. It's part of my identity. It was my thing," musician and audio producer Mark Henry Phillips said in a recent US public radio feature, explaining how AI seemed superhuman, excelling across every musical genre, style and instrument. "My special skill just isn't that special anymore. From a musical and economic point of view, AI just has me beat."

The difference for Omarji has to do with her idea of creativity. She describes herself as a "big user" of AI for tasks like idea generation or quickly reading and summarizing documents. Yet she doesn't use it in her music. "For me, music has always been about self-expression and creativity, and those are not words that I really associate with AI."

"I want to keep [music] a human process," she adds.

AI-generated music: Its own sort of art?

Pasquier also doesn't believe that AI systems are creative. "They are imitating the data they were trained on and lack intentionality and framing," he says. "They are missing quite a lot of what makes human [and] artistic creativity." 

But, he adds: "Making a machine that makes art is a practice in itself, known as generative art, or metacreation. This is not new and has been a niche with its followers for a long time." 

Ultimately, however, music is more than just sound, AI-generated or not.  

"People like an artist, say a musician, for their look, for their attitude and personality. There is a lot more to fandom than the music itself!" says Pasquier. "Now some systems are starting to introduce these characteristics, and this uneasiness becomes excitement to some." 

But such excitement eludes Sophia. She sees how AI could give musicians a pedestal who have something to offer yet might not fit certain industry standards. But engaging with music for her is largely about discovering the artist and their story. She's skeptical that an AI artist can deliver this.

"If [a song] is created by an AI, then you go in and see that they don't really have a story," she says. "And I think that really takes away from a big part of what the music industry is today."

Edited by: Stuart Braun