Right: Dr. Nettrice Gaskin's Deep Dream-driven portrait of Kendrick Lamar -- AI-based art that preceded Midjourney and other generative AI platforms.
Generative AI platforms like Stable Diffusion and Midjourney have been available for general use for about 2-3 years, and in terms of content created, we're seeing a cascade of imagery that mostly fit in these broad categories:
- Decoration: Enhancements of existing images, as with this avatar realification project I mentioned recently.
- Illustration: Pretty/dramatic images depicting a specific subject or theme. Basically, the kind of images you'd see, say, on the front of a record album or book cover, or to advertise a game. Like this image that a judge subsequently refused to allow the generator to claim a copyright on.
- Mash-Up Memes: Leveraging the fact that Midjourney and other platforms have vacuumed up decades of existing popular culture, these are usually fun mash-up pictures/video of the "X meets Y" variety, like Star Wars meets The Simpsons, or Barbie meets Oppenheimer.
To be clear, these are all entertaining and occasionally valuable uses of gen AI (assuming various lawsuits by artists are resolved.)
But there's one category of imagery that seems conspicuously missing: AI-based fine art.
In other words, imagery which expresses something essential about the human condition. I.E., images that you'd imagine seeing in a regular wing of the Museum of Modern Art, not just a special one-off show about art and AI.
Where are they? It's quite possible I've missed some notable works, but I follow this space pretty closely.
Frankly, the only example I've personally come across are works by Afrofuturist Dr. Nettrice Gaskins. Years before generative AI platforms even existed, she was creating powerful images of Black artists using Deep Dream as a palette.
Dr. Gaskins sees her work as a continuation or an extension of hiphop culture. As she's written:
The nature of the tool is to find new styles, new aesthetics in between existing ones, esp. for those of us who create or make images. Generative AI feels like sampling and remixing, two techniques from hip hop music production that I have some experience with or knowledge of. These techniques were part of a culture that was synonymous with grit, resilience and creativity.
The images she creates are beautiful (see above) -- but they also have something relevant and important to say. As metaverse artist Jeff "AM Radio" Berg put it to me, about her work:
Although beautiful, the work puts the learning bias of race to the forefront, compounded by systemic racism’s affect on the digital divide and unequal access to advance tech tooling such as machine learning. In the work, suddenly it’s no longer about ML, but about the story of our culture and society in its interactions with technology.
Realistically, systemic racism means that most people of color will encounter machine learning in the face of racial profiling by machine learning algorithms. If it’s during a protest or a simple traffic stop, this is the reality for many, ML and data in general is a source of fear. This inexcusable disparity of experience is highlighted by this far different experience of Afrofuturist possibilities that exist despite the bias.
So her works speak to artificial intelligence as a subject of interrogation of the artwork itself. This also seems to be theme explored by Stephanie Dinkins, another African-American artist now featured at the Guggenheim. (And someone Dr. Gaskins pointed me to.) So at the very least, we have two artists using AI as a medium working on a high conceptual level -- whose very subject is AI in relation to race and power.
But these seem to be some of the very few (only?) examples of AI-based art that's very arguably art in itself -- and they were both working in the medium before Midjourney etc. But where is all the actual art being generated by Midjourney etc?
I guess one answer is that it's still early days. But then, I thought these platforms were supposed to accelerate human creativity. And they've already been around for some two years!
Or maybe the better explanation is generative AI simply is not, generally speaking, a good tool for fine art.
Unless I'm missing many more great examples?
Comments