Here’s a lovely interview with Dave McKean on Gil Roth’s Virtual Memories Show.
The context here is that McKean has created a limited edition graphic novel called Prompt, composed entirely of artwork he has created by feeding text prompts into Midjourney, a text-based art generator AI that is currently in vogue, particularly on social media.
There’s a lot to love about this interview. For one, I like that McKean tried the thing before passing judgement on it, and read up on it (the book he mentions – The Artist in the Machine – sounds fascinating in itself), and has tried to think about various angles on this, which means there’s a lot to chew on for the listener here. McKean is deeply concerned with the ethics of art, and it’s a fantastic filter for the questions thrown up by AI-generated art.
One of the things he brings up is the rise of Photoshop, which devalued some of the craft of photography and illustration, but which, of course, is a tool rather than a generator. The idea here feels similar to the rise of digital typography, for example, which means that more of us can create basic typography much more easily, but of course, that doesn’t mean that we’re all able to do it at the same skill level, even though of course the result was to put a particular kind of craftsperson entirely out of work.
The depressing thing about this kind of AI generator is, of course, the fact that it’s not just making the work easier for you, it’s actually doing the work.
This feels to me far more similar to a lot of modern conceptual art, where the “artist” comes up with the idea, and then feeds it through the labour of others (artisans, or in fact other artists), yet claims the credit of creation. The question of ownership and in fact creatorship gets far more interesting here.
This line of thinking can be continued into ideas of appropriation and plagiarism that the art world has been (or has been refusing to) grapple with.
Midjourney and Dall-E can create, based on text prompts, art that looks like it was created by any artist that you might care to name. So then, is that work of art “created” by you, or by them, or have you, in some way, stolen their style and the work they put into creating it?
We can tie this back to questions that the comics world has faced for a long time, from the time Lichtenstein took a panel of Russ Heath’s artwork for his own, or when Sharon Moody did photorealistic paintings of existing comics pages and exhibited them.1 In both of these cases, these artists at least put in the physical work of replicating the original. AI art takes that out of the equation, as well as a lot of the abstractive tendencies of each artist that we tend to refer to as “style”.
So what would this mean for the art world, if someone would, like McKean, do the work of generating a prompt and curating the results and call it art? If something can be entirely yours even if several others participated in its creation, why would this not be “yours” in the same way?
I don’t have specific thoughts in this direction, but it doesn’t feel like a coincidence that we’re grappling with this and the digital scarcity/ownership ideas that NFTs throw up.
Of course, at the moment, while this AI art is “good” (for a certain value of good), it’s not yet good enough to take over the entirety of the art context. However, it’s not too early enough to start thinking about the fundamental question this throws up – what is “art” if there’s no human intervention? What is “art” if, to take this further, it’s not made for humans?
It’s an interesting sf idea, perhaps, to have the art world integrate art from AIs, and ideas about how alien these intelligences might be and how they might read us even as we make them “be artists” for us. McKean talks about Solaris at one point, where the horrors are merely an attempt by an alien sentience to get in touch. We assume that an intelligence generated by us might be understood by us, but that is in no way a given. How might this kind of art function if we make it a conduit between AI and us?
The final question that I found interesting in what McKean says is probably his biggest one in practical terms. He mentions that The Artist in the Machine is extremely concerned with how AI might make art, but it doesn’t really ask why. This has been a failure of humanity for a very long time – we spend so much time thinking about what we can make and how, that we don’t ask if we should make certain things at all.
At one point, McKean asks if we actually need shitty AI-generated poetry? The question becomes a little more interesting when we consider the idea that AI might actually make good poetry. Would we need that poetry?
What do we need poetry and art for? What do we need stories for? And is it essential to us that a human should be on the other side of it?
We can look at CGI and VFX in this context. A lot of CGI work done on environments and backgrounds is procedurally generated with AI. Some of it is hand-animated. The former is generally understood to make things easier for the creative artist and get a lot of low-priority things out of the way so they can focus on the latter, which is the more “artistic” component.
I think we tend to be happy with automation and replication when it comes to the same kind of low-priority elements, as long as there’s a human being looking at the high-level stuff and making decisions about it.2
But what if the same thing could do it all for us? Would that “art” be satisfying? And again, as I was asking above, would it even be “art”?
Anyway, just thinking out loud. It’s not like we’re going to get answers immediately, and it wouldn’t be entirely clear if those answers even change anything out in the real world. Just interesting to think about.
- I also remember that the Icelandic artist Erró repurposed art by Brian Bolland, Chris Weston, and Guillem March, used them in collages, and exhibited and printed them. However, I can’t seem to find any related images. Here, however, is an open letter written to him by Brian Bolland. ↩︎
- This is an interesting question to look at in comics lettering, where we’re still wondering if we lost something and what that might be in the transition from hand-lettering, where someone is literally drawing every letter, to digital lettering, where we copy-paste text in typefaces and move on to work on everything else. ↩︎