Back in April, I wrote an article about the wonders of art created through artificial intelligence. Programs like Dream by WOMBO, Starry.ai, DALL-E/DALL-E 2, and Midjourney are dominating the AI art scene right now. Even at just four months old, that article is wildly outdated according to what AI is doing these days. And stick around for the last section of this article, where we talk more about the one that we made ourselves using AI.
If the idea of computers creating visually stunning images scares you, just wait until you see what AI artists have been making lately.
A TikTok User Shared Their AI Results
TikTok user Robotoverloards has been showcasing what they think are the most interesting pieces of AI art, and a recent post has really caught the attention of the internet.
Simply using the prompt “selfie at the end of the world,” AI artists have conjured up some truly terrifying images. And I’m not really sure what’s more horrifying, the fact that AI knows what scares us the most, or the fact that these images are compiled from real, existing images on the internet.
@robotoverloards Asking an Ai to show the last selfie ever taken #creepy#scary#ai#midjourney#dalle2 #foryoupage ♬ It’s Just a Burning Memory – The Caretaker
If you’re not familiar with how AI art works, it basically uses advanced algorithms, fed by text prompts entered by the user, to take hundreds or thousands of pieces of public images to build a final image that may or may not make sense. You can specify everything from the exact placement of your subject on the image, the final product’s aspect ratio, the art style, main color palettes, etc. So if you wanted to see what it would look like if Salvador Dali painted the Mona Lisa, you could that. It might be a little wonky, but that’s art, my friend!
And since the AI uses existing images, you have to know the right prompts to navigate the AI’s limitations. For example, human faces and general facial alignment is the most difficult thing for these AI programs to master, but more advanced applications (like DALL-E 2 and Midjourney) are improving this feature daily.
@robotoverloards Asking an Ai to show the last selfie ever taken Part: 2 #horror#scary#ai#dalle2 #midjourney #foryoupage ♬ It’s Just a Burning Memory – The Caretaker
AI Created a Last Image Photo for Us Too
The header image on this article (and displayed again immediately above) was made by me in Midjourney with that same simple prompt, “selfie at the end of the world.”
I didn’t specify any other details. I didn’t say anything about skulls or drab colors or anything else, and that’s what the AI produced.
But even if I put that prompt in again, it would come up with another version that’s not identical. That’s the beauty of this art: every piece is unique and the rights to these images belong to the creator.
Midjourney is a Discord-based application while DALL-E 2 is its own thing. DALL-E 2 currently has a waiting list for its beta that includes over one million eager users.
As I explained in my previous story, AI artwork “involves computers using a convolutional neural network to learn and create entirely new visuals from enhanced patterns in existing images.”
Google created its own AI art tool called DeepDream in 2015. Now we have DALL-E and Midjourney and more. The results are intriguing and also terrifying.
Want to chat about all things post-apocalyptic? Join our Discord server here. You can also follow us by email here, on Facebook, or Twitter. Oh, and TikTok, too!
