Art, AI, and Social Positioning
"It can create, but it cannot perform."
SIDEARM: AI, Art and Social Positioning
It’s raining.
I’ve got my camera bag, my backpack, some lights. Standing in the rain, trying to figure out what counts as the ‘front’ of a set of rows out by Strathearn Heights--someplace left of centre in Edmonton, Alberta. Most of where I write from stands somewhere in this city.
The screen of my phone is too wet to ask my interview source, Riley Hornbeams, to let me in (yes, that last name is indeed an alias. I think). Riley is a less-than-stranger from the internet: a staunch communist with which I’ve had many (MANY) pen-palled conversations about the nature and purpose of AI. He’s talked a lot about artists who train their own AI models to produce art. He’s my first stop of the day.
I thought of those folk - - these cellular communities of AI artists that Riley described - - in a sort of bemused way. Kind of tongues the punk art collectives of the 70’s in my imagination. It made me think of roomfuls of people who want to challenge the idea of what art can be, which tools are ‘allowed’ -- which tools ‘count’ -- as Art. Then I remembered lessons about the history of the photograph, the composite image, the sale of objectivity, the emergence of the hyperreal, the synthetic ideal, the death of the artist, the ‘death’ of painting. I think of the hundreds of posts in the aether of the internet - - artists railing against the use of AI generation, tools, and products.
I think of the fear, too. I thought about people carrying anxieties that their careers are automated into oblivion. I thought of that classic productivity VS profits chart that is pulled into almost every conversation about modernity, economics, the working class. I think about the areas in our daily lives in which we invite automation, and the areas in which we don’t, and the band of difference between who invites what, the Montessori education system.
In Riley’s view, artists (particularly visual artists) are angry at a symptom of late-stage capitalism, which is displaced onto AI as a set of creative tools. He also argues that it’s mostly a diffuse, vague, and reactive movement. “Bad artists,” he called them (whatever that means), who he feels lack their own identity, style, financial security, and etc. Riley argues that it’s mostly that group that’s genuinely afraid of AI-generation in art, not because of its (and their) artistic ‘value,’ but because their identities are wrapped up in the scarcity and the gatekeeping of art. He also fundamentally disagrees with the very idea of intellectual property, and emphatically describes himself as “not really an artist,” so, take that for what you will, with salt.
We spoke a bit more and I kept thinking about photography, the problem of the composite image. But are composite images not art and those who create them not artists? I guess that’s a debate that’s still ongoing - - one that positions art as most valuable in its interpretation across contexts rather than the work of its creation. Then there’s the argument that the creative process determines soul/soulessness, but does that detract artistic value? Or add to it?
These are, of course, arguments not best had, made, defended, and explored, by two people (relatively) well-fed in a warm basement somewhere.
By two people not standing out in the rain.
Overall, I left our conversation with the lingering, tempered feeling that, perhaps, AI in a vacuum is not necessarily the enemy of creatives and art. Not on its own. But I also left with more questions than answers. Questions like “is art for everyone?” and, again, fundamentally what is “art”?
There many arguments for accessibility to the craft of art. There are also arguments for the craft of art (the Doing or the technical) being inextricable from the artifact (the product, the experience) itself. So does more accessibility to craft or process detract from the value of art (both cultural and economic)? Well, maybe. It really depends who’s at the helm.
The role of AI also calls attention to the role of the artist or author in the interpretation of the artifact -- what the art ‘does.’ If the author doesn’t matter, then all composite art is validated, and thus AI-generated works have interpretive value. At the same time, things like Stuart Hall’s cultural studies look at how representation in arts and media tacitly reproduces 5-dollar words like ‘hegemony.’ So, what tacit structures and meanings does mass web-scraping also reproduce to be realized across contexts? And, economically, does this further democratize art? Or does it just reproduce the problem of concentration? I’d argue that, unless checked now, it absolutely does.
That’s where this gets into the weeds.
And I needed to touch grass.
I needed to get out of the basement and ask an actual artist.
In researching how to make this piece topical, I was pointed toward the dispute between Kendrick Lamar and Drake. Theirs was a decades-old beef that turned red hot around the end of this past April. As of this sentence, it’s May 13th, 2024 (and the release will likely extend into June at this rate).
The thing that got my mind churning again, circling back to AI and society and our next steps as a species, was Drake’s use of unauthorized vocal likeness of Snoop and Tupac in ‘Taylor Made.’
Shakur’s estate was extremely f*ckin’ pissed at their necromantic monopoly being threatened and instantly moved to file. Snoop was gob-smacked, and announced this on his own socials, despite the fact that he’s no stranger to licencing his voice and likeness in exactly this way. What was actually shocking to me is how ubiquitous AI voice generators actually are.
Snoop, Drake, Lamar, Shakur -- if you google any famous person’s name, there are dozens and dozens (and dozens) of generators available, free of charge. But when I think of ‘free,’ I think of the saying by someone much smarter than me: “If you aren’t buying it, and you aren’t selling it, you’re the product being sold.” Then I suspiciously eyed the freshly self-installed Windows Copilot bouncing for attention in the bottom-right of the screen, a symbol of that creeping ubiquity.
Ding. (The discord sound. You know the one. Well, not really a ding, but I couldn’t onamatopize it in time for this deadline.)
It’s my friend Shayne, at CKUA, trying to point me in the right direction.
“Try Arlo Maverick,” they said.
“OK.” I said back.
And then I thought about sampling, too - - the practice and tradition of sampling and lifting in hip hop. And I thought about where that new line might fit into the climate where nothing feels very real anymore.
Nothing.
It’s no small feat for a true and utter nobody (yours truly) to book after-hours studio space in YEG. For some reason, I’d convinced myself it was easy. I wanted to respect the time and space of anyone who agreed to the semi-sane idea of having a stranger on the internet (yours truly) point a camera at them and ask about the nature of art, reality, intelligence, consciousness, the duality of humankind as it grapples with artifice. Shit like that.
After about a week of emailing back and forth with Arlo’s manager, Natalie (bless her), we came to a solution. I finally get to meet someone who’s got something to say. Someone who creates things, well-situated in an industry standing at the precipice of AI in art — specifically music, this time. I grabbed my friend Syr to be my second camera (thank you, by the way) and off we went.
Arlo oozes awareness and intelligence, and, to be frank, coolness. Even when thrust into a “who the hell is this guy” situation (me, I’m the guy), he struck me immediately as the kind of person who’s always thinking, and when you pry him on something, he’s able to boil out the meaning behind it very quickly. A true creative.
“I think that the mixed reaction to it shows that we're not sure where we're going with this,” he said. “I think that we are it's still too early to be afraid, but we do need to be mindful of it.”
Like the switch from horses to the steam engine, or the automobile, Arlo described, also, that the tools themselves weren’t the problem, and that there are elements of art that AI can’t touch.
There’s the product, the process, and the performance of art. AI deals mostly with the product.
“It can create, but cannot perform.”
This is (somewhat) true. If art is fundamentally a relational act (not to be confused with the overlapping relational aesthetic), then AI cannot replace artists. The experiential element of one human creating a piece of themselves that exists as an artifact, but also as a moment for other humans to experience.
That’s what art is, right?
When I make a piece of me that sees a piece of you.
This answers the question, ‘who is art for?’ A simple answer is that it’s both for the artist and the audience. Creation for creation’s sake. The value of a moment of exchange that is, as Arlo described, “something magical that [takes] place within that moment of those human beings being there and all these different perspectives coming together.”
A piece of me that sees a piece of you.
This still doesn’t quite answer to what extent AI should be involved in the production of art, but Arlo - - with HUGE SPOONFULS of caution - - suggests that it can be a tool on the toolbelt. We just have to be cautious on how we wear it. If people choose to use it, instead of using it to skip the process, we’d ought to use it more like an instrument. Naturally, we’d got to talking about the practice of sampling and how it compares:
“Sensitive subject, because of the fact that you're taking someone else's art and repurposing it in a way that, if done right, creates a whole new creation. If not, just seems like you're just blatantly copying something with no respect for the creator of it
[...]
In the 80s, when hip hop artists were sampling, a lot of these producers, didn't have access to instruments, didn't have access to studios to record this stuff.
[...]
If we fast forward to now, when we look at how that has changed how other genres of music have become more accepting to it, because for some it's because, okay, now there's contracts and things in place to make sure that everyone is fairly compensated.”
And that’s the crux of the issue. All these tech companies rushing to the bottom to scrape as much content from the internet as possible to train their models without compensation or attribution before the legislation could catch up. But things are starting to change - - just as they did around sampling technologies in the 80’s and 90’s.
It seems like AI is here to stay. As will its involvement in the creation of art, or content, or both. Arlo distinguishes the two - - as do others. One is a compromise over the other RE: the financial realities artists grapple with. And while AI certain lends itself heavily (and bleakly) to the creation of the latter, its involvement in the arts is almost certainly not going anywhere. As a tool on our toolbelts, it’s up to us, as creatives, to decide how to use it.
AI and its place in the future rely on human intentionality.
Syr and I left that interview feeling a whole lot less replaceable, and I think we all gave each other a lot to think about.
My favourite kind of day.
Arlo had given me the number of yet another person much cooler than me - - someone who had specific ideas and opinions about AI and using it in their work. Someone who I would later find out had a lot to say about the importance of intentionality.
Aristotolese Canga.
I texted him immediately, and he got back to me the next morning.
Ari is a producer, an engineer, and a musician rolled into one. Arlo sent me his way specifically because Ari has experimented with AI - - though not necessarily with his music. Things like cover art, however, were a different story.
“It’s a really good thing,” Ari said. “It complements, an artist's lacks.”
Now, this isn’t to say that Ari is dusting his hands and saying he’s OK with AI displacing creatives from the field. Far from it - - also fundamentally believes that AI is just another tool in the toolkit and that it should stay that way, for the purpose of augmenting creative reach or addressing issues of accessibility:
“I see a lot of African kids that don't have access to the the resources that we have, but they do have a computer and they have OpenAI. They can type in a prompt and they they can generate art from that. So it's and from the result of that, those prompts, we can get access to the product of their imagination as well.
[...]
I would still consider it a product of someone's mind and someone's imagination, which is the aspect of AI that I like.”
Intentionality.
And when I asked Ari specifically about that, he paused and thought for a full moment. Then he smiled.
“I would just mimic the words of a really famous musician that I like. His name is Wynton Marsalis, a trumpet player. He said the thing that AI doesn't have - - and he even goes further in saying that the thing that I will never have is intentionality.
[…]
My favorite songs, the songs that I really like, they have these imperfections in them, but those are sometimes choices of the person that is working on the song.”
To Ari, AI-generated songs are just “nice.” Not much more. They check the boxes, but they’re missing a human element that’s hard to describe.
The relationality, the improvisation, the moments of creation: so far, he feels AI cannot create what is so precisely imperfect and great about all art - - especially music. AI cannot have intentionality. AI cannot build relationships.
“Where does art come from? Where does the song come from? We don't know that. Like, is there a higher power that feeds us these thoughts? AI doesn't have access to something like that, something greater.”
I think of Arlo performing in front of a crowd, mic to his lips and hand outstretched.
I think of Ari in the studio, one headphone on, smiling over a keyboard to a guitarist, a bassist, a vocalist. Or maybe no one at all.
Then I think of Ntwali Kayijaho, driving a red convertible through familiar Edmonton streets - - his music video, ‘A Message to the Youth.’
My producer gets us in touch (thanks Alex).
Ntwali’s main gripe with AI, as an artist, is how creators - - especially new creators - - seem to lean on it a bit too heavily. As said earlier, art is the product, the process, and the performance. It’s as much about improving yourself through the development of the craft, and the product contains those stories.
It’s about “the living aspect,” said Ntwali:
“Acquiring a large enough palette to write music, the human aspect of waking up every day to go to a job, or struggling academically, or struggling to have self-belief or self-confidence. Working your way up from the bottom.”
AI is a short-cut, he feels. And people rushing for name-tag - - the status of Artist™ - - means that people risk missing out on that lifelong process of self-improvement, self-discovery, and self-refinement. They are missing out on the connection that Arlo and Ari talked about, too. But the promise of AI is too glossy for some people, and Ntwali fell for a similar ideal early on in his career:
“That's one thing that like, I've learned with music. When I first started making music, I thought I was going to make a hit song when I was 18. You know?
But that’s not how it works. Sure, you might have the hype for it, but you got to learn the fundamentals and you can't cheat the steps.”
It’s ongoing, Ntwali said, and the longer that he’s doing it, the more he sees the importance of not getting distracted by the glossy promises of shortcuts to fame and fortune. He keeps his eyes on the real goal: “to impact lives and inspire people.”
Ntwali is a bit inflexible in his stance on AI. At best, he described it as “a gift and a curse,” citing a case or two where singers have lost their voices and used AI to recreate it. So, again, honourable uses for AI. But at the same time, there are cases on the horizon of AI being used to steal voice-actors’ likeness too.
A gift and a curse.
He’s inflexible, but said something that made me laugh (out loud) when I talked to him about the practice of sampling, comparing it to AI scraping.
“All great artists steal.”
Yes. All great artists steal.
And this is precisely why process and performance are so much more important than product, and why art, as a concept, is much bigger than what we see. When I heard Ntwali say “all great artists steal,” it all clicked for me: a crucial difference between sampling and AI scraping.
AI has given us a good set of tool, a GREAT set of tools. It can fish things up for you, create a composite work that follows prompts drawn up from your imagination. But you might not know what they mean beyond their affect. The context, the histories - - the process, the intent, and relationality - - might be lost in the use and reuse of these symbols on a surface level.
Without the historical, relational, or experiential elements linked to the craft, process, and network of art, the value and purpose of criss-crossing sampling lineages can become meaningless. Muddled.
The indexical orders of meaning created through sampling in hiphop is a core value to the scene today, and that value is at risk of being lost. Each song is a sign, and when that sign is moved, it brings with it a set of contexts and meanings; it accumulates knowledge and meaning as it’s handled, moulded, experienced, and rearticulated.
Generative AI carries the risk of overwriting it, convuluting that context. You can argue that you’re a wikipedia article away from ‘knowing’ the histories of these samples, artists, and music, but there’s also the emotive knowledge. The experiential struggle and relationships that is shared across generations - - is that something that can be replaced? And, here’s a harder question: does that NEED to matter to an artist for AI-generated stuff to ‘count’ as art?
I guess that depends if we value the product of art, or if we value the process or performance.
Arlo, Ari, and Ntwali all seem to value the latter more than the former.
And I do too. Because why do we create the product in the first place? Essentially, what is art?
ARLO:
“There's something to be said about artist development. There's something to be said about having those moments where you fail in the moment, and then you have to go back to the drawing board and figure things out.
One of the reasons why we as human beings have been able to create the things that we've created is the fact that we've been we've been given the opportunity to fail. We fail. We have to do it, to innovate something and create something that is going to help us overcome.
And in doing so, ideas are born that would not be born if we always had the cheat code to just get us to where we need to be.”
ARI:
“Art is nature. It's nature. Everything around us. It's life.”
NTWALI:
“Working on yourself. Trying to become the best version of yourself. These are all things that make us human: learning from our mistakes and trying to elevate and trying to grow.
I think those are the things that makes true artists.”
And AI can’t replace that.
Not yet, anyway.
Not yet.
Whether or not that stays true is really up to us.
-zd