Join Puck to listen to this article
It’s been nearly seven years since Christie’s first sold a portrait produced by A.I. at auction, achieving more than 60 times the work’s $7,000 estimate and bringing in more money than an Andy Warhol print and a Roy Lichtenstein bronze work offered in the same sale. At the time, the house declared itself “the first auction house to offer an artwork created by an algorithm,” and wondered whether A.I. was “set to become art’s next medium.” Since then, they’ve continued to test the market—just a few months ago, they dedicated an entire auction to A.I.-generated art, and well surpassed their estimates to bring in more than $700,000.
The adoption of A.I. is accelerating across nearly every creative industry. But it’s worth asking: Why should A.I. be doing anything creative in the first place? After all, generative A.I. is effectively statistics, and statistics can be seen as the opposite of creativity. And yet, as my colleague Ian Krietzberg points out, art might be especially vulnerable to disruption—the technology may have a tendency to “hallucinate,” but that’s hardly an impediment in this field.
Earlier this week, we discussed what gets lost when A.I. attempts to create art, the trade-off between speed and quality, where the technology might actually help the creative process, and much more. Our conversation has been lightly edited for clarity.
“It Doesn’t Matter if It’s Good”
Marion Maneker: Why is it that people involved in A.I., and in Silicon Valley more broadly, have this passion for eliminating originality or artistry, the human creator? I’m not opposed to A.I. in any way, but what I don’t understand is why it would be applied to art. Is it marketing? Is it for proof of how powerful it could be? Where does all this come from?
Ian Krietzberg: You’re right on the money there. It’s kind of a mix between marketing and proof of capability. The funny reality is that so many prominent folks in this industry will always talk about the holy grails of the pursuit of this technology, which is like, Cure cancer, solve climate change. It’s ironic to me because very little of what they’re doing is actually pursuing that. What’s turned out to be the case is that, when you’re dealing with these generative systems, the easiest way to show you’ve mastered some form of capability is in the arts.
These systems aren’t reliable. They have all these caveats, questions, and concerns around them. What they can easily prove is, you ask ChatGPT to write you a book, it’ll write you a book. It doesn’t matter if it’s good. They’re able to prove a very generic capability that gives the impression of general application. Art is a place where the mess-ups, the hallucinations, and reliability don’t really matter.
There has been an enormous growth of bullshit over the last 20 years with the amplification of these platforms, and now we’ve invented a technology that generates bullshit at higher rates.
Because of the environment we’re in, where attention spans are shorter and the key is more content, what the owners of studios and media companies want is more content, faster, and cheaper. Those are the three things that I keep hearing: more, faster, cheaper. No one is talking about better; no one is talking about more artistic.
And I’m not sure what that gets you, except confusion and a rising tide of stuff that your customers have to wade through and feel beleaguered by.
The thinking from some of these places is, If we can just fill a pool with stuff for people to swim in, maybe they won’t get out of the pool. The money is no longer in a single TV show—it’s in keeping people subscribed to the platform. And how do you keep people subscribed to the platform? More stuff, all the time. That’s how the game has shifted. I don’t think that’s necessarily the right approach. I feel like I’ve been hearing from everyone for years that there’s just too much stuff. For a while, there was too much good stuff. Now, there’s just too much shit.
Bruce Forever
In a previous interview, you were talking about your interest in authentic art. Talk to me a little bit about your sense of authenticity versus A.I. as the technology expands.
Everyone loves Bruce Springsteen. If you listen to a Bruce song, it’s not intensely complicated. When he sings, it may not be the perfect pitch or the most technically correct, but the grit in his voice feels very emotive to me. If you sing about something painful, which he often does, you really feel that. That’s what’s lacking to me.
Art isn’t about perfection or being technically correct. There are so many musicians who know very little about music. John Mayer can’t read music, but he has a feel for it—he’s fluent in the language of the guitar. Not everyone is Billy Joel when it comes to technical capability, but they’re communicating something about themselves in a raw, impassioned way. And that’s the element you can’t find in a synthetic machine.
I think there’s one aspect of the whole entertainment complex that is the opposite of A.I., and that’s sports. The connection between sports and the authenticity of art is that it’s human-based and has an unknown, unpredictable outcome. I feel that with art, music, and, to some extent, storytelling writ large, there’s still the opportunity for very talented people with original ideas to do things that will attract attention and move the ball forward.
I think for a lot of sectors, media included, what’s happening here presents a massive opportunity. What we’re likely to see with generative A.I. as it gets adopted and pushed everywhere is that it’s just an averaging down of everything—an averaging down of thought, creativity, and consumption. And average isn’t interesting. So if everything gets pushed to this bare line that’s tailored based on your cookies and preferences and such, that creates quite a significant opportunity for weirdness.
