Todd Dobbs’s captivating journey through AI-generated imagery and its complex relationship with human perception—packaged in a witty exploration of art and technology—challenges assumptions about the “typical American.”
Picture this: it’s the year 2022, and artificial intelligence image generation is all the rage. Our protagonist, Todd Dobbs, Denver-based freelance photographer and associate professor at the Rocky Mountain College of Art + Design in Lakewood, Colorado, was already knee-deep in the non-fungible token project Art Block when he stumbled upon Midjourney AI thanks to a server invite from an NFT artist buddy. Now, Midjourney offers Dobbs a monthly AI service, akin to those early cell phone plans where unused minutes go *poof* into the digital abyss. Determined to not let those precious minutes slip away, Dobbs delves into a realm of stream-of-consciousness prompts.
And guess what? He strikes creative gold with one prompt that sets this whole adventure in motion: Photograph of a Typical American. Yet-to-be-explored characters are waiting to be uncovered, and I’m sure there’s more to the collection of disturbing and mordacious Coca-Cola-addicted human eagle mashups on the horizon.
The images eloquently unveil AI’s take on Americana, painting a picture of a time when burgers are celebrated with more gusto than a Broadway show. We’re talking about juicy patties that sizzle like they’re auditioning for a cooking reality TV series, the air redolent with the promise of freshly baked buns and gluttony.
Guns, in these images, aren’t just accessories; they’re like prized family heirlooms, reminding us of a history when muskets met Manifest Destiny.
And the American flag? It’s not waving; it’s practically tap dancing, bedazzled with a 1950s flair. It’s an emblem of hypernationalism that’s been hitting the gym for centuries.
But beneath the glossy surface, there’s a sneaky undercurrent, a subtle rebel yell of apathy, cleverly masked with a smooth Tennessee whiskey finish. It’s like the AI is saying, “America, you’re a complicated masterpiece. Let’s raise a glass to your shade.”
Now, why that phrase, you ask? Well, Dobbs has a knack for exploring subjective phrases that are open to a zillion interpretations. He’s on a mission to stoke debates and discussions among us humans, and he knows that AI relies on our word associations to get the creative juices flowing. The project’s charm? It’s all about identifying trends and themes in AI-generated interpretations of this seemingly innocuous phrase.
The creative process? It’s a mixed bag. While Dobbs’s prompt choice is the real star of the show, the technical bits follow. Dobbs engages in a one-on-one convo with the AI bot, where he gets a sneak peek at the AI-generated images. Then, it’s all about the upscale. No filters, no cherry-picking—every single option gets the deluxe treatment. Dobbs has a strong catalogue game, neatly sorting results based on shared traits for some top-notch research.
But here’s where it gets spicy: the dynamic between AI interpretation and human perception is an often contentious struggle. Dobbs knows AI isn’t sipping tea and discussing the meaning of life. Instead, the algorithms, data, and consumer-driven advertising materials used as sample imagery generate visual amalgamations.
Oh, and did we mention the number 1776? That was Dobbs’s chosen endpoint, which he blew past quicker than anticipated. That’s the number that, in combination with images that roast Americana, irreverently lampoons a pivotal year in American history. It’s when the nation tentatively began its journey, attempting to unravel the complexities of what being “American” could mean—somewhat akin to a perpetual identity crisis, marred by its colonial and imperialist history. Quite the drama, isn’t it?
Here’s the bottom line: AI is a trusty sidekick, not a replacement for human creativity. It’s like the fancy new tool in your toolbox that makes tedious tasks a breeze. AI might remix, but it doesn’t birth entirely new styles. The collaboration between humans and AI? It’s like brainstorming on steroids, influencing traditional art forms along the way. What’s fascinating about Dobbs’s prompt is the results that AI is spitting out quicker than you can say “glitch.” Though he’s been using this prompt in multiple iterations of Midjourney, the images all fit into a clear visual collection with little shifting in color tone, facial apathy, and style.
But it’s not all about Dobbs. You’re part of the narrative, too! Your interpretation matters.
Dobbs’s hope? To kick-start discussions about the mighty power of keywords in the AI realm. Those subjective words hold more sway than you’d think. Dobbs aims for discussions that leapfrog political divides, focusing on what it means to be a “typical American,” in a chance to challenge assumptions and stereotypes.
The artist’s collaboration with AI has rekindled connections with long-lost acquaintances, sparking conversations that stretch into the wee hours. But one reaction that really got the gears turning? A former student pointed out a lack of diversity in the results. Dobbs took note and started paying closer attention. The results? The images are becoming more diverse, slowly—but still unmistakably polarizing, reflecting the evolving AI landscape and its lack of ability to truly showcase the diversity of the Typical American.
So what’s next for Dobbs and his grand project? Generating, sorting, and researching. Narrative projects might also be on the horizon.
“The first goal is to figure out how to display these for gallery view so that every Typical American can be seen at the same time so the viewer can be completely immersed in that world,” says Dobbs.
Nothing is stopping the public from running the exact same prompt on their own, so if you’re up for an interesting afternoon of AI collaboration, feel free to jump in. This not only makes the project more collaborative, it truly makes it part of the public domain.
Dobbs’s parting wisdom? Consider how keywords shape the AI world. With AI taking center stage in our lives, understanding how these systems interpret keywords is vital.
“This aspect of computer learning will not go away in our lifetimes, so just like a teacher has to adapt to different learning styles, now we must adapt to how keywords influence the results that a [generative adversarial network] produces, and humans who contribute to training it—even if involuntarily by just posting images on the internet that it eventually catalogues,” Dobbs explains.
“We know what is in the image and what it should mean, but we have to explain that to an algorithm using language it can understand now,” he says. “I also hope viewers contemplate what being an American means to them and acknowledge some of the false assumptions we make when that term is used to describe someone.”