Is it okay for AI to make music?

Hey everybody, I know this is a surprisingly relevant topic for me to blog about but I’ve been thinking about it since my last blog post so here we are.

In my last post, one of the youtube channels I recommended was this guy Universal ambients. Now the channel says that all of the music on it is produced by the guy who runs it but, it did get me thinking about AI and music. Especially non lyrical or ambient music.

At first look it seems like ambient would be the perfect type of music to be “AI’ified” but there is of course I think a general sentiment against AI producing any type of art, even if it’s advertised that AI was involved. This, however raises the question of what music even is. I’ve heard of some people listening to sorting algorithms to study, so would it be wrong to get an AI to make something similar sounding an then marketing it as a sort of ambient noise? would there be a difference if I marketed it as ambient music?

I think in our time now, we don’t really have to worry about this because the primary factor in why we listen to music is how good we think it is, and AI music is not good so we don’t have to worry about it. In the future though, I could see it becoming a bigger problem.

Also, it seems like now, and even in the future, there’s really no reason to listen to AI music over human made stuff. I mean, there’s not a lack of ambient music out there, If I want to listen to music, wouldn’t it make sense to listen to human made stuff just for the sake of it being music? I think so.

Anyways, those are just a few thoughts about AI music to round out the year. Let me know if you guys have any opinions on this.

5 responses

  1. The connection to sorting algorithms is an interesting one, since it is an automated process. Yet somehow, I cant image AI music sounding anything like sorting algorithms. Based on “art” by LLMs that we’ve seen up to this point, at least, I would imagine it sounding increasingly similar to regular old human music, just without the personal some of the artistic flavor that we’re used to.

  2. Wyatt Haas Avatar
    Wyatt Haas

    I completely agree with your statement that there’s no reason to listen to AI music over human music. I don’t see why anyone would disagree with that. I do think that, unfortunately, it will reach a point where AI-assisted music will become a lot more common, especially when it comes to saving money within these cash-cow labels. The only place I can imagine fully generated music being a commonplace occurrence is through corporate music, i.e., elevator music. It’s nothing new, but I can see AI make significant strides in this department, just look at the Muzak company in the 90s, there is a demand for it. One side note that I thought was cool to mention was how the Muzak company plays music that encourages work, which means the music speeds up slowly over the course of 30 minutes and then suddenly stops. Pretty bizarre, but kinda cool, and I think AI could do this without a problem.

  3. Clara Yeaman Avatar
    Clara Yeaman

    This reminded me of a video I watched that talked about AI in kids media on youtube. There are a good amount of Kids videos where there’s a song and a little animation, at least one of which is generated by AI. Older people may judge music based on if it’s “good,” but kids can’t really make that kind of determination. I wonder if or how that changes the conversation.

  4. An interesting line of thought you could branch off of from this is the technological capabilities of AI, especially relating to creative human endeavors. We know that AI is especially effective at completing tasks with set goals and objectives with set data, and an example would be telling AI to sort a spread sheet or some income statement by expenses or whatever. AI would destroy these types of tasks, because they are supercomputers capable of processing at speeds far beyond human ability.

    This begs the question, however, when will AI be capable of accomplishing subjective, human-like tasks such as art, music, or other creative tasks? The very things that make us human and we hold so dearly like music are not made with a list of necessary components, and are forms of creative expression. AI is not yet developed enough to simulate human expression to the point where it would be indistinguishable from human work.

    Think of all the AI-generated artwork that’s being pumped out in mass quantities. It is clear in the AI-generated images that AI heavily struggles with replicating human text, or human faces, or fingers. A seemingly easy task, yet AI fails. Now imagine AI trying to create music with rhythm and beats and sounds that engage our human brains? That seems like a huge milestone if AI was to be able to replicate it effectively.

  5. Prof L Avatar
    Prof L

    Last semester we were talking about a poem in my class, and I mentioned in passing that a line from the song–“stay in this world with me”–sounded like the chorus from a pop song. Before the end of class (50 minute session) a kid raised his hand and informed us that he’d used AI to generate a pop song featuring that line! (And it sounded like a legit, mediocre pop song….)

    Re. Miles’s final big-picture comment, and also in relation to Clara’s comment, I think that things will get complicated soon, once AI music starts influencing humans who go on to make music influenced by AI music,,,

Leave a Reply to Wyatt Haas Cancel reply

Your email address will not be published. Required fields are marked *