Marc Ribot leans over his guitar in his performance at INNtöne Jazzfestival 2024. His white hair os frizzy and he looks at a fellow musician over his glasses.

Noise, art, and AI. A musician considers the function of art

Marc Ribot responds to a question on whether AI can create.


So, one of my first experiences with ChatGPT was I asked it to write a poem. Okay, I said, write a poem about a palm tree.

And it scared the heck out of me. I hate to say this; I get goosebumps saying this: It was a decent poem. Okay, now I know that it generated that on the back of real poets who sweated it to put out their stuff. But getting away from the issue of the unfairness—which is manifest, right, and the need to reimburse—am I just being sentimental when I say that it kind of gives me the creeps that that AI could write what I thought was a good poem. Or do you think that artistically—especially if we could deal with the giving credit where it’s due—do you think that art could ever be created by AI that you would listen to or see, that you would love, that would move you? Do you think we’re we’re coming to that, or do you think it’s always going to be anodyne? Or do you think that it will always be identifiable by an artist.

I think it’s possible….

Right. I want to go one further gloss on one further approach to your question because it’s a good question. So, the real answer is I don’t know. I do believe that, you know, x number of monkeys typing in a room for x number of years could create Hamlet’s soliloquy on death. And I would like it. It would be good, you know, so the answer to your question is Yes, I do believe it’s possible that AI could create something.

But what about the question of whether it would serve the function of art. It is a very deep question.…

There was a former French minister of the economy, Jacques Attali, who wrote a book called Noise.1 And in it he cited an anthropological theory that concert music derived from rituals of human sacrifice—that early societies had problems with unchanneled violence and they created violent rituals in order to channel the violence to control it, to limit it. This was, believe it or not, beneficial. These rituals were accompanied by music, and in time the music replaced the actual killing, the actual violence. (Um, except in pro football.) But it still had to serve the function of channeling violence, which is why every new iteration of music—Papa Haydn’s “Surprise Symphony” [for example]—was going to destroy society. Elvis Presley was going to destroy society. Louis Armstrong. Swing bands. They were all going to destroy society. Punk rock, you know. Gangster rap. Every new iteration of music is perceived as horrible, threatening, dangerous, malevolent because it’s channeling social violence. It has to serve its function.

Now, the people who do that, in order to do that … to do that trick of finding the next thing, the thing that’s going to do the trick for young people who need music, you know, that’s going to channel the current forces, the current social unresolved social forces into a new art, that’s going to channel our violence, the violence inherent in human society—that cannot be done by a machine that by definition only looks at the past, only synthesizes the past. Or maybe it can be done again. Monkeys, you know, monkeys with typewriters, but the odds are pretty low.

And so, it’s not: Could it write a good poem? But: Can it write the poem that we need that we’re going to die if we don’t get?

I don’t believe it can.


Marc Ribot talked with Tift Merritt and Misha Angrist in mid-March 2026 in a meeting of Duke’s Science and Society “Periodic Tables.” The topic was AI and music and musicians. Ribot and Merritt are performers, and also leaders in a growing movement to organize musicians.

The question that Ribot considers in this excerpt from the discussion was whether AI can create music and poetry, and Ribot’s answer is nuanced. He thinks AI will get “better,” a judgment that Tift Merritt agrees with, but he argues that the truer question for musicians is whether its products—what I like to call music-like sounds—fulfill roles that music and poetry play in society. Ribot thinks AI will not measure up: The question isn’t “Could it write a good poem?” but: “Can it write the poem that we need that we’re going to die if we don’t get?”

Music serves many purposes, and the emerging purpose that AI-made music-like sounds serves seems to me to be close to what Muzak once served. Muzak was simply commercially produced “background music” that quickly became known as “elevator music,” because that’s often where people heard it. But the purveyors of Muzak services claimed that their product could increase sales in department stores, lift morale and productivity among office workers, and even was assumed to play a role in profitability of bordellos. Joseph Lanza reports in Elevator Music that “workers employed in the world’s oldest profession at a brothel in Stuttgart, Germany, grew concerned because their uptempo Muzak ‘Light Industrial’ selections were not getting customers in and out fast enough to secure a profitable turnover. The proprietors had to make a special request for livelier music on their second and fourth quarter-hours.”

AI “slop” is derivative and, as Ribot notes, is “anodyne” and backward-looking. It is yet another twist on the strange story of music and technology. The role of music has changed along with technologies of listening: Spotify automatically offers background music for our lives, and even has playlists that fit the moment of, say, taking a shower in the morning. The phonograph changed where people listened and, indeed, who played the music for whom. Sony’s Walkman turned music listening into a solitary affair. Streaming services changed the economics and availability of music, with profound effect on the relationship of music listeners and their musical choices—if indeed algorithmically selected music count s as a human choice. AI music-like sounds are pushing music economies and musical experience yet again.


  1. Jacques Attali, Noise: The Political Economy of Music (French title: Bruits: essai sur l’economie politique de la musique) Trans. Brian Massumi. Theory and History of Literature, Vol 16. University of Minnesota Press, 1985. Alibris: https://www.alibris.com/Noise-The-Political-Economy-of-Music-Volume-16-Jacques-Attali/book/53636761 ↩︎

Image credit: Schorle, Deutsch:  Marc Ribot Jazz Bins Auf Dem INNtöne Jazzfestival 2024., July 21, 2024, Own work, https://commons.wikimedia.org/wiki/File:Marc_Ribot_Jazz_Bins_INNt%C3%B6ne_2024_09.jpg. Rights: CC Attribution (4.0).

See also