šµ Everyone calm down! A pragmatic look at AI and music
As lawsuits fly, let's try to take a balanced look at this all...
Hi everyone,
Luck is running against me this week; I am writing this on day 3 of Covid, which a good number of you might recognise as āpast the night sweats and onto the persistent headache and coughingā phase. A laugh a minute, Iām sure youāll agree.
For that reason, this is a cut-down āop-ed intro onlyā edition of Network Notes today, though at some point I am keen to do a quick reader survey to establish which parts of NN people actually enjoy and connect with, if only so that I can make the process of delivering this to you all the more valuable. But for now, letās dive inā¦
One would have to have been under a rock this week to have missed the news that the RIAA is suing both Suno and Udio. The reasons are obvious enough to anyone that has used either platform: at this juncture, only a madman would dispute the notion that both were trained on masses of copyrighted content.
You need proof? No problem: when someone manages to get a near-perfect recreation of Mariah Careyās āAll I Want For Christmas Is Youā track (click here for a sample courtesy of the WeaverBeats channel - most others have now been scrubbed), you know that 1) your defence is likely going to crumble, and 2) you will have most definitely poked the bear when youāre messing with one of Sonyās highest-earning tracks.
Of course, the issue here is not that these platforms are capable of spitting out copies of songs. The problem is that the rights holders arenāt being paid. Solve that - as will surely happen now - and the landscape might well be permanently changed.
So is this it? Are we now doomed to a slow descent into AI-generated hell?
Iām not so convinced. Thereās no question AI will challenge things on all manner of levels, but a recurring theme I am seeing anytime I look at the intersection of AI and creativity is that we are increasingly coming to value what it is to be human, and consequently what contributions humans make to the creation of art.
Previously, this was a mildly absurd notion. Granted the evolution of the synth might have prompted similar concerns (e.g. Kraftwerk and ārobot musicā), but on the whole it feels like AI is now allowing us to recognise human effort all the more, and place an even-higher value on it.
Personally I still feel AIās contribution to creativity should not be underestimated either. Many platforms are popping up that Iād argue empower artists and creators all the more. This is not a binary matter of āAI = badā. As someone who makes music too (and whose company works with a number of AI platforms in the music creation space), I am loving the possibilities and potential here. Why? Because all the examples I see empower artists further.
However, at times, I feel that the AI industrial complex might just overplay its hand, and in doing so, may just invite something of a backlash. Again, this is the heart of the issue for me: AI can most definitely do some great thingsā¦. but it cannot do anything or everything, and that is the root of the problem. I could deliver a very, very long list of things AI is really helping with at present; the issue is when it oversteps the mark.
I remember watching a Grayson Perry documentary and him commenting that all art is essentially imperfection. Every style, nuance and variation was more often than not borne of the artistās own inability to perfectly recreate something. That is where the originality creeps in, because often those imperfections are like a fingerprint; not something you can easily clone. Once created, Iām sure someone can try and have a go using AI (try asking any AI image platform to make you a Basquiat-inspired piece and youāll see my point), but those first pieces came from who-knows-where in the artistās mind, and that is the true value here.
Perhaps the irony is that the louder the din of AIās hype becomes, the greater our passion for all things human becomes.
The story, the mythos, the broader aspects to art are also factors that draw us in. Creation of the artefact was never enough; there needs to be a story around that and reasons for people to connect with it. And connections are human. Context is everything.
Today Iāve been lying in bed, listening to Diamond Jubilee, the new album from Cindy Lee. It is only available via two means: on Youtube, or via a Geocities website that looks like something straight from 2004. No Bandcamp, no DSPs, nothing. And yet, here is Fantano, raving about it, and here is The Signal (another Stack I highly recommend) citing it as one of their albums of 2024 already. At a time of everyone clamouring to tell you that you must do this thing or that to succeed, seeing an entirely outsider prospect like this catch so much love is a sure sign that we are not even close to being subsumed into some AI-powered matrix of shitty songs. And I am fairly sure that one factor in it generating so much interest is that broader context of it being an outlier, not on any music platform besides YouTube.
Another argument of AI is that it might simply replace the easily replicated churn music that lacks any originality. Massive Attack put it best:
This is why that Cindy Lee album landed so well with me. It is quite bizarre at points; like a weird pop fever dream led by a band that may or may not be able to play their instruments in time together. Itās all part of the charm - all part of the imperfection that draws you in.
So yes, there is a lot going on right now around the AI debate, but personally I feel that art, culture and human connection will always endure. It might evolve, but it always has, and when it has, weāve often been the better for it.
For now, ironically, I feel like the more AI tries to tell us it can be our culture, the more we all come to reject that concept.
Isnāt that reason enough to be hopeful?
Have a great weekend,
D.
Get well soon, Darren!