Does typing prompts into an app make you an artist?

Our columnist generated this image in Midjourney with the following prompts: "A male and female model kissing, dressed in Gucci beach wear while lying in the Great Plastic Patch in the middle of the ocean. Fashion photography. Highly detailed. Vibrant colours." And no, he doesn't claim it's arty photography. — JASON GODFREY

This might be a hot take, but artificial intelligence-generated images are not art.

Let me get into this. For anyone who doesn’t know, Midjourney, Dall-E, Stable Diffusion and others are artificial intelligence-based (AI) apps that can generate images from just a few words, which are called prompts.

Want to see a man walking in a city? Type that in. Want to see a cartoon version of a man walking in a city? Just add "cartoon". In short, anything you type, the AI – which was trained on billions of publicly available images – will imagine it as an image, and produce it in seconds.

This has spawned a lot of Twitter account-holders posting their AI-generated images while also posting their prompts. This is particularly helpful for anyone generating images as it gives you an idea of what can be done and how to get there.

Late last month (April 2023), one of these Twitter account-holders that posts images and their prompts, dropped the premise of posting their images under the guise of being instructional, asking if they could post their creations “prompt-free” – I can only conclude because they’re proud of their “art”.

And here is where I will come at it again and say AI-generated art isn’t art. But that doesn’t mean it can’t be.

AI-generated art has won awards. A man from Colorado, United States, generated a digital painting called Theatre D’opera Spatial that won first prize an annual art competition. Though it did win in the digital category and right at the start of the entire AI-generated anything craze. When he posted his win online it received a lot of backlash, being called “the end of artists” or just plain “gross”.

OK, so it won in the digital category as I said, but I have to agree that it seems unfair that something that is created in seconds like that can be called art. It seems like an insult to artists who train for years to get the discipline and skill necessary to create something similar by hand.

But the prompts, people argue, that is where the real talent comes in. The ability to tell the AI in a clear and concise way to put together the elements you want to see, surely that takes creativity?

And this is true, to imagine things no one else has gives AI-gene-rated images a novelty. Reddit threads showing “what if bears where friendly” with a whole thread of hikers posing for selfies with bears, or Harry Potter as a Wes Anderson movie, are a lot of fun to look at. But is this art?

It’s imaginative, yes. It’s fun as a Reddit thread, but again, it’s not art. Writing a prompt, no matter how complicated someone wants to make it, doesn’t yield better results all the time.

As someone who has generated a lot of images on Midjourney for work and for fun, sometimes the more you type, the worse it is, and the less you type, the more beautiful the image. The AI is adept at making something beautiful because it’s trained on what we humans consider beautiful.

Here’s where we get back to that Harry Potter as a Wes Anderson movie – those are great fun to look at but the AI is lifting from Harry Potter and Wes Anderson movies to do it. It’s mashing up two artistic pieces of work, it’s not creating anything new, and you for typing the prompt definitely aren’t either.

Generative AI is trained on the backs of artistic genius, and the images it creates are beautiful and gorgeously weird and insanely bizarre, but you as someone who writes prompts for the AI are not an artist. Sorry.

It is amazing to sit and create images that match your mood. I spent an hour generating images of a boy and his dog throughout time and space, and realised it was because I was feeling a bit melancholy. And on that level generative AI was giving me a release, I suppose, just like creating art would – but it took very little effort and because of that the resulting images were nice, but lacked weight. I haven’t looked at them since.

I’m not saying someday we won’t create art using generative AI. Humans are definitely, crazily inventive and will find a way to create true art with AI, I’m sure. Just don’t type in “totally awesome sunset” and then show it to all your friends like you’re a photographer, because you’re not.

---

Big Smile, No Teeth columnist Jason Godfrey – a model who once was told to give the camera a ‘big smile, no teeth’ – has worked internationally for two decades in fashion and continues to work in dramas, documentaries, and lifestyle programming. Write to him at [e-mail] and follow him on Instagram @bigsmilenoteeth and facebook.com/bigsmilenoteeth. The views expressed here are entirely the writer's own.