I asked AI image generators to draw me. Here’s what came up.

by Gordon Haber

Artificial Intelligence is like death in that it’s inevitable, and there’s nothing we can do about it, so we may as well prepare—especially if you earn your living from editing and writing. 

Most recently, I’ve been tooling around with image generators. Not because I relish the idea of illustrators or photographers getting put out of business. But it does seem important, as a freelance writer, to understand the limitations and attractions of the technology.

First I played around with Stable Diffusion. Actually it was stabledifffusion.com—note the extra “f”—because that’s what comes up when you search for “stable diffusion” with DuckDuckGo. 

After trying various prompts, out of boredom or vanity I typed in “Draw a picture of Gordon Haber.” Here’s what came up.

Here’s the metadata by the way:

“prompt”:”Create an artistic portrait of Gordon Haber, a wildlife biologist renowned for his work with wolves in the Alaska wilderness. Depict him standing amidst a lush, green forest, surrounded by towering trees adorned with vibrantly colored foliage. Utilize soft, golden sunlight filtering through the leaves, casting gentle shadows on the ground. Capture the essence of his passion for nature with a thoughtful demeanor, wearing practical outdoor attire that reflects his adventurous spirit. The scene should exude a sense of tranquility and reverence for the natural world, illustrating the unique bond between humans and wildlife.\n\ndraw a picture of Gordon Haber”,”originalPrompt”:”draw a picture of Gordon Haber”

Aside from the fact that the AI generated its own prompt, which is problematic, it also confused me with the wolf expert named Gordon Haber, who doesn’t look like the picture above. He’s also dead, but that’s beside the point.

Here’s what stabledifffusion.com came up with after I cleared my cache and tried again.

I’m not sure what to make of that bizarre disembodied hand. Maybe it’s a reference to The Hand, the 1981 movie where Michael Caine plays an illustrator whose hand gets cut off and then crawls around strangling people. I am a big Michael Caine fan.

Next I tried openart.ai with the same prompt. It gave me two versions. In the first, I look like Kevin Spacey’s disappointed dad…

…and an antebellum state senator with a hangover.

I wanted to try an image generator from a more reputable company, so I tried Microsoft Designer, which provided four images. First, it rendered me as a young, Asian female…

And a young, Asian male…

And an over-caffeinated white guy…

And again, another young, Asian female.

As surprising as this was, ChatGPT was arguably the most interesting of the experiments. Here is the response to my prompt:

While it is indeed flattering to have my work referred to as “witty” and “observant,” I’m not sure how I feel about being rendered as a depressed-looking 40-something with several rejections from N+1.

I’m also not sure if there are any grand conclusions to draw from this experiment. Except maybe in the case of Microsoft Designer, either the algorithm sucks or there are a surprising number of Asian women named “Gordon.”

Finally, in case you’re curious, you can see a picture of me on my LinkedIn profile.