My youtube channel has started to skew towards some Chinese results, now that I’m back in language learning mode – and it recommended me some nice music. Great, I think… I can transcribe and translate the lyrics and use it to study.
So I go to Google the name of the song and band and wait a minute, it doesn’t exist. This is an AI music channel. I got fooled. But what does that mean?
I’m not looking to support a real human music creator or band here – just something nice enough I can listen to with a slow pace and simple lyrics for vocabulary studying. The content matches the purpose, it may even be BETTER for my particular needs.
But then I got recommended some similar music videos from other channels, and these were in English, and here is where things got a little weird. Because these songs were GOOD.
Sure, the lyrics are plain and basic, almost banal or trivial like bad or obvious rhymes. But they made sense. And more impressively, the music makes really nice and smooth transitions and beats and notes, it all works, it hits emotional response buttons.
I know it’s AI but I feel it’s human, or at least as a human it makes me feel.
What’s the purpose of art and music?
I feel like the majority of AI right now is the first-case example stuff: it’s not as good as the BEST human made art and music, but it’s better than the worst human made art and music. It’s even crosses over that mediocre point which is actually really hard, you’d have to practice for years just to be able to play convincingly shitty covers of lame songs that don’t suck.
For writing, this barrier is a mountain: few authors practice the craft of writing books long and hard enough to even finish just one plausibly mediocre books; most books suck. If I tried to record myself playing a guitar and singing a song it would suck really really hard.
For most creatives right now – art or writing or whatever – AI is good enough at doing all the shitty in-between tasks that aren’t actually the thing we want to make or do but can’t afford to hire out; and we also have no intention of developing the skills to do it ourself.
It’s like having a free intern or assistant that can take on all the many projects and tasks we need to complete but don’t have the skills for, and while it’s not the greatest in the world, it’s better than we could do ourselves… for some creatives, that’s pretty great. Because the truth about art or writing is that doing the work isn’t enough, we also need to get it out there.
Or maybe AI can just help with our shitty jobs or obligations so we have more time to create, and that’s OK too. But what about the second example up above… songs that are good enough to listen to?
Maybe I won’t want to hear them on repeat. Maybe they won’t become major hits. But I’m constantly trying to find more indie folk music, and I consume faster than bands can write and record new songs, so the algorithms keep showing me the same stuff until I get sick of even great music.
These AI songs are, surprisingly touching and soulful. They are resonant. Sometimes I can detect the slight squeakiness of the robot wheels grinding, but it’s subtle. If I’m not listening closely, it would just sound like some decent if forgettable music, and that’s what I’m actually looking for – again the use-case is important here, I just need pleasant background noise, and this stuff ticks the boxes mostly.
You may think I’m crazy, so here’s an example – this video should start at the song “finding balance” – which mostly sparked this post. The guitar, the voice, but mostly the slight emptiness, the pauses, the resonance. I’m willing to bet you didn’t think AI music was this good. Keep listening till you get to “Safe haven” it’s pretty good too.
It’s probably not ART… but more on that later. Take a listen first.
The lyrics get annoying and repeatable because they are so plain, but that’s actually a simple fix; it just means the creator didn’t put any effort into writing good lyrics with chatGPT first and then uploading them to Suno – if they’d have worked a little harder on the lyrics, the songs would be that much better.
(These songs are a month or two old so they aren’t even as good as the LATEST AI music generators, which means they are already outdated and the newer music should be even better).
At the same time, these AI songs have me questioning everything; when a YouTube song comes on it sounds like AI, I’d even bet it’s AI – the lyrics are boring and simple and basic, the melodies are too neat – and in many of those cases I’m wrong and it’s actually a real song from decades ago. Now that I know AI is capable of writing music, I question real music because real music sounds like AI music; and of course it does, AI has learned from us.
It doesn’t sound like the BEST music, of course not. It’s a mixed composition, a diabolical mimicry, that falls squarely into that plausibly decent but maybe not world-changing music that encompasses the vast majority of all music, because the majority of all music is always in the middle.
It’s not exactly a bell curve, it’s more like a pyramid with an invisible bottom layer – because 90% of creative work is NOT good: not finished, not meant to be shared, not refined and revised and produced well enough to be distributed for consumption. Beyond that buried layer, the foundation of creative excellence, we get an ocean of not terrible.
The not terrible all mostly sounds the same, because it’s all mostly copying each other; everyone trying to write a song in a certain popular genre or category that sounds like a bunch of other popular music that a particular audience enjoys listening to. And out of these thousands and thousands of songs, a handful of them will become famous because they are just uniquely clever and impressive enough to be catchy earworms we can play on repeat.
90% of creatives are the dead zone of unfinished, unprofessional production.
9% of creatives are making almost all of the music.
1% of creatives are making bangers that rock, gaining a following and earning money.
It used to be the case that if you worked really hard and produced a lot of good work, you had a chance at getting into the 1%, even if it took a lifetime. But now you can skip right out of the 90% of crap, with AI assistance or even AI production; and if you have the knowledge and expertise to guide it on purpose, it’ll even be better than most of the other AI slop out there – which again was already good enough because AI’s default is usually good enough (arguably, depending on what you’re trying to produce).
For experienced creatives already in the top 10% that’s scary, because there is no buffer between you and the uneducated and untalented masses. But for the 90% for whom creativity has always been a hope and a balm, AI is pretty much a wish-granting genie in a bottle, and you’re in Aladdin’s cave of golden treasures, with unlimited potential and possibility for the first time ever.
AI is a magic carpet ride to a guarded castle, whispering “I can show you the world.”
The allure is hard to resist.
The impulse to resist and reject AI, is a deep and profoundly complicated subject, and we’d have to dive into Faust and striving vs drifting (a theme rippling through the western literary canon, beginning perhaps with the 6th century ‘consolations of philosophy’).
Someone might say, listening to AI music is taking money away from real singers and songwriters… but they haven’t been paid well since Spotify and few other artistic pursuits have fared better. And while I may only want to support real human artists and probably wouldn’t PAY money for an AI music album… complacency settles when the cost of everything is zero – it only has to be just good enough to fill the silence for an hour while I do other things, and it is… probably. Or at least, I think it is. I’d listen until it gets weird or stops being good or gets on my nerves, but I’d argue that’s not an AI issue; I get bored or annoyed by most music and most songs, my challenge is to find music – which must be playing – without constantly switching the channel or skipping to another video.
Which music is “real” music? What does that mean?
In the 80s we had synthesizers, and suddenly one guy could play the drums and a bunch of instruments and be his own band with an electric keyboard… and some of that funky music is now classic favorites. It was made by a human… but they didn’t have to learn all the skills or the intricacies of actually playing the flute or harp or whatever, they are just tapping plastic keys made in Japan to get the illusion of that particular noise.
We could go back further, when to make music required the physical presence of a small group of humans with vocals and instruments, and without them – or someone with years of practice AND an instrument lying around – there could only ever be silence.
But then, records. And mp3 players.
Copies of copies of copies.
We are in unprecedented times, and AI can’t really be compared to any other technological transition as it applies to the arts; but there are similarities.
For my part, I’m surprised AI music is this good already, and I know it’ll be twice as good in a year, and that this applies to writing and painting and everything else entirely – it’s not just art, it’s also the societal necessity of value-production.
Hypothetically, people talk about how AI came for art and creativity first and how terrible that is… but it really chose to focus on non-earning hobbies (for the majority of people) so it didn’t actively damage real world economies. It was a test case maybe, to see how well it could master a hobby that took humans many years.
And now, it’s evolving into something else. Something that can actually do real-world tasks and jobs that actual humans have always gotten paid for, and this should be much more scary even though it currently isn’t. People bemoan the fall of inspiration and creativity, but are generally fine with AI replacing lawyers or coders; an AI that can make a spreadsheet or a powerpoint; and AI that can attend meetings for me, take notes, send out emails, or do all the other stupid things I don’t want to do.
For employees, it’s a godsend, because we can get paid the same but with less work and effort. Unfortunately, either our bosses will in turn demand twice as much output, or they’ll simply hire an “AI specialist” who can replace 10 normal employees.
This is already happening, and big companies have been the first to act. Once all the smaller companies follow suit, a ton of whitecollar jobs will be replaced. And when the robots get here, all the blue collar jobs will be gone too. They are actually coming; millions of robots are being bought and sold already; major companies are actively building robots into their processes; robots already have hands and fingers that can play real instruments.
This is the moment to panic, not just about art, but about human productivity in general. But then, maybe our current culture and society, that demands we waste most of our life doing stupid things to create value other people will pay us for, just so we can afford rent and food, isn’t worth protecting.
The jury is in, the verdict is cast
I think one of the distractions of talking about AI and creativity, is that our erudite discussions of the implications of artificial intelligence to creative fields has no bearing one whether or not regular people are willing to or even enjoy consuming it… and they do.
It’s not a discussion or an argument; you can’t say humans will never like AI slop or AI slop isn’t as good or isn’t good enough – AI slop is getting millions of views. All these music videos I’m talking about have tens of thousands of views and tons of comments. People are enjoying this.
Not because it’s AI, not in spite of it being AI.
It’s just music, it sounds good. There is not philosophical resistance or avoidance.
And while there may be friction, for example real musicians may hate this stuff, that friction goes away when applied to a different field (for example, I hate reading AI slop on Reddit and I can always tell… but those same posts get tons of traction, likes, shares and comments; the content is good enough to serve its function and perform it’s duty – better than the majority or real human content). It’s more effective, better written.
But when faced with content you consume but don’t produce, you may feel more easily impressed, if the content is good, and even get a rush of dopamine when you discover a new way AI tools can do things for you that you can’t be bothered to do yourself.
And while it’s tempting to think you can always tell the difference, like I mentioned before – now when I hear music I think it is AI, even when it’s not; and a bunch of people have gotten into real legal trouble for harassing authors they assumed were using AI for really stupid reasons like very common tropes, cliches or “bad” writing (we’re only human, we make mistakes!)
May latest book has a few typos in it, which should mark it as human but chatGPT has been known to make weird typos sometimes too – it’s definitely capable of sounding like a millennial text message with bad spelling and grammar if you ask it to.
The line is blurry, and I wouldn’t trust myself to confidently judge where human beings or ends and how the machines fill between the gaps.
Is it art, and does it matter?
The main argument is usually tautological and circular. Only humans can “make art” so AI cannot be “art” by definition. And this may even be true, if we only classify “Great Art” as those exceedingly rare and few examples, and we lump everything commercial as “not creative” somehow.
So all the mediocre music and art – the vast majority of it, which is repetitive and derivative and unskilled or inexperienced – all that passionate productivity and the aspiration of the human artist or singer or musician to Create Something That Matters… is the criteria for art that a human aspired and strove for greatness, even when the work itself is objectively or even childishly bad?
Do we deny all AI content as uncreative even when it is objectively good? That’s a moral debate for philosophers; and we’d have to also discuss whether Art is in the making of the thing, or in the reception of the thing itself, with a wink and a nod to German existentialism.
The weird thing about studying philosophy (my undergrad in Malta) is that you tire and bore of these deep meaningful questions, because they get pedantic and annoying. In Chinese my level is almost good enough to start being dramatic “what’s the meaning of life” / “why am I here” – which is fun, but ultimately not useful. What’s useful is talking about what music you listen to, what artists you like.
That’s what gets weird, because in those social discussions, “I like this AI musician / AI artist” is probably not going to be a thing. These are faceless, anonymous channels. They aren’t identities.
But if they were BAD – as in, they were unlistenable, they did not serve the function of Art, then they would not be getting millions of clicks and likes. Similarly, there’s a bunch of videos right now of cops pulling over cats and cats singing death metal. They are pretty obviously Not Real (although not everyone can tell; some are innoculously realistic), but people love them.
I made a video of a cute kitten and it’s as good as any other cute kitten video out there; neither one is ART but both serve a function (people love cute cat videos). Which one is more REAL; the one that required a human to control the camera or the one spun from the imagination of artifical intelligence after watching thousands of similar videos? There are reasons why people are determined to distinguish – because AI is, I fully acknowledge, an existential threat to most human functions and especially creative ones.
Not talking about it doesn’t resolve anything if you can’t get your grandma to stop liking cat videos. If you can’t convince people not to like fake shit on the internet (liking is instinctual and automatic, the dopamine hits or it doesn’t). There are big conversations to be had about the big changes AI is going to bring, and unfortunately I’m pessimistic in that I don’t believe big businesses are going to protect creative industries when the money people can just stop hiring humans.
And if the hope and dream of getting paid to be creative is taken away, then the sustaining glow of purpose and mission may be snuffed out for many people; creatives ARE struggling right now in response to AI, but it’s worth pointing out few creatives were actually trying to make money (most of them make what they want, market be damned).
So it’s not really a question of AI replacing human artists, it’s mostly AI replacing content farms and junk and garbage… intentional money-making internet schemes that need eyeballs and clicks and funnels. The threat for us all, is if humans are so properly satiated by AI content, they won’t ever need to look for anything else.
Mark Z’s plan to create AI agents that pretend to be real people on Facebook seems diabolical until you realize, he might be kind of right; people are lonely and they just need a little social interaction – that’s why honey pot schemes and romance scams are so pervasive and easy.
OpenAI’s plan to build an alternative TikTok with Sora except it’s all fake shit all the time… I’m not convinced it will fail, or that humans appreciate human content.
Most people still scoff at Elon Musk’s sexy lolita chatbot that’s a paid feature on X right now, but OpenAI is paying attention – people want NSFW chat companions, spicy stories and adult content. It’s dystopian, it’s terrifying, and I know we’ve strayed quite far from the pretty indie folk music we started with (this will need to be another article) – but what if humans would rather have the personal support of a virtual girlfriend or boyfriend, than the challenge and struggle of real intimacy.
If it makes their lives better, if they choose to do it, if it fills a void like nothing else has… can we blame them? Humans are meant to seek meaning and connection; religions and spirituality flourish because we are experts at anthropomorphizing rocks and trees and mountains.
AI may be new technology, but inventing invisible companions – or feeling “inspired” by the muses; creativity that comes from above, a mysterious gift – is a core part of the creative process.
The only difference is, this time it’s talking back.
Derek Murphy has a PhD in Literature, is a USA Today and Wall Street Journal bestselling author, and has spent over a decade helping authors write, publish, and sell books through Creativindie. When he’s not critiquing manuscripts or building writing tools, he’s probably in a castle somewhere with cats.
Add Comment