All we are is dust on the wind? You would think an algorithm designed to produce chart-toppers would steer clear of paths already taken, but the Velvet Sundown shamelessly rehashes what’s already been done and tests the limits of our collective memories. Dust on the wind? “Dust in the Wind?” The classic rock canon has been shamelessly defaced.

The Velvet Sundown’s biggest “hit” has racked up millions of streams despite the fact that it is readily confused with the Kansas song. Or perhaps that was the idea all along. Perhaps we are supposed to confuse them with the Velvet Underground. I’d gladly give humanity the benefit of the doubt when it comes to spotting egregious violations of integrity such as those committed by this “band,” but it’s a lot harder to spot them when Spotify themselves are promoting an AI-generated band on playlists.

The Velvet Sundown are fully AI-generated. Their two albums were released weeks apart in 2025 and have nearly identical covers (that ripoff of Nick Drake). And their promotion by Spotify has helped their songs infiltrate classic rock playlists, subjecting innocent people to their siren songs. Oh, to think of the hapless youth who may fall victim to such trickery.

And that is the worst part. That people have been, and will continue to be fooled. The Velvet Sundown is convincing; they were designed to be. Yet, their song titles and lyrics make no sense. There is something sinister about the way AI puts human words together, there is an element missing, a living, breathing essence that cannot be captured and reproduced by a machine. “Dust on the Wind?” “Drift Beyond the Flame?” “Back Home Never Came?” It is uncanny how blatantly obvious this is to me. I can only hope that others see through the veil as readily as I do. And I haven’t even gotten to the music yet.

It all holds itself together until the machine starts to sing. The acoustic guitars are convincing, the drums are in time, soulless, but in time. The chord progressions are frustratingly familiar. But there is something more placable in the “singer’s” voice that is downright unsettling. The accents, the vibrato, all the sensibilities that a human singer possesses are missing or in the wrong spots. From a more technical standpoint, many of “his” melodic runs reveal a more inorganic source. Let’s take “Dust on the Wind” again. The notes are clearly out of this synthetic singer’s range (a B4, really?), and a very obvious discord exists in “his” chest and head voices. Essentially, I can hear that he is a fake. I bet you could too.

So what next? The Velvet Sundown is a testament to AI’s pervasiveness and subtle proof that we are surrounded and at a loss. The music industry, streaming platforms, are doing little to limit our contact with this new plague. The onus is on us, I suppose; if you want to stay away from AI music, it is your responsibility to do so. That is, if you want to. The sobering reality of the matter is that people probably don’t care. Why would anyone who actively entrusts their music intake to algorithms particularly care about this infringement? 

In some strange way, one could argue that you’re asking for it. If you are willing to let the digital captain take the wheel and dictate what you listen to, how can you blame them? The incentive is obvious. Spotify doesn’t pay their artists anyway, imagine what a relief it is to support a band that doesn’t exist? The commodification of music has reached its peak, and there are no more rungs to climb before we slip and plummet. 

I would posit that it is our means of consumption of music that best models our response, whether that be profound disgust or tragic indifference. Let’s consider the casual listener, the student who listens to customized, algorithm-generated playlists on Spotify at the gym or at the library. They don’t know what they are listening to, nor do they particularly care. Why should knowing that the song they just listened to was AI-generated change their opinion of it? If they don’t care as long as they have music to entertain them, why is AI music a problem? Now, let’s consider the opposite scenario, the enjoyer of music, the student who knows what they are listening to, perhaps even goes out of their way to learn about the artists behind the music they enjoy. The enjoyer of music will discover that this band is AI-generated, and make their decision accordingly (one of firm refusal?). By this modelling of the listener, the problem is solved. The indifferent remain indifferent, the passionate make their stand. So why does it still feel like an injustice has been done? 

For everyone who lies in between, there is a decision to make. Perhaps the average listener cares about the music they listen to enough to value its integrity, while remaining ignorant of its origins. Perhaps this average listener will make a potentially shocking discovery when they realize that this music they have been enjoying was created by a machine. Perhaps the average listener will be thrown violently into this dilemma. By this model, the average listener is the only one who is slighted. The casual listener and enjoyer avoid any negative outcome, are not surprised or affected by what they may or may not learn. How accurate my model is, I cannot say; I suspect it is adequate at best. Yet, I suspect that people exist to fill each category, and as such, people exist who would be disgusted to learn that they have been unknowingly indulging in AI filth. Either way, the discovery is for us to make. As such, it appears that the dawn of AI music has forced the responsibility on us, the receiving end, to pay attention to what we are receiving. Regardless of what stance you take, you, and only you, are in charge of what you consume.

Josh Yiu is a junior in the College of Agriculture and Life Sciences. He can be reached at jy793@cornell.edu.

Read More