I Tried an AI Image Generator: Things Got a Little Weird

I should have stayed away from animals

Artificial Intelligence generated images are having a moment right now. Curious to see if I could create an artful masterpiece by simply entering text into a box, I decided to check out the popular AI program, DALL-E2.

According to the website, “DALL·E 2 is a new AI system that can create realistic images and art from a description in natural language.”

DALL-E is available by invitation only. I signed up for the waitlist and received an email stating, “to ensure responsible use and a great experience, we’ll be sending invites gradually over time.”

Eager to get started creating my “artwork” and not in the mood to wait, I did a quick search and found a free online AI program that lets users create a limited number of photos per day.

NightCafe is an easy-to-use AI art generator that can create images from simple prompts.

I signed up for an account and received seven free credits to use on my first day.

Image #1: Pug at the beach

I entered the text prompt, “Pug at the beach watching the sunset.” And, behold, my masterpiece.

What did I get? I got a giant and sickly-looking dog, lying on its side and gazing somewhat lazily at what appears to be the silhouette of a naked person sitting up in purplish water.

It’s nightmare-inducing and not anything I would ever consider hanging on my wall. I’m pretty sure that dog is considering whether or not it's going to reach over and eat that floating person. And when it does, it won’t even get up to do so. It will just lie there and chew.

Image #2: Another pug

This one was a user error on my part. After inputting the text for the first image, I was prompted to create an account. Not realizing the first pug image had already been created, I attempted to enter the same unfortunate text again. Happily, I found I was given a new creation. Unhappily, I think it may be worse than the first one.

The zombie dog’s face is distorted, it’s missing an eye and part of its face. And what on earth is that standing above it? I see feet, but a leg that goes nowhere and some odd shape jutting off to the side. A flipper maybe? Whatever it is, I want no part of it.

Image #3: Log cabin in the forest

Not wanting to see any more dogs, I decided on a simple prompt of a log cabin in the forest and chose the cubism style. This one is fine, and a bit of a relief after the first two. I kind of enjoy its simplicity. Although it’s a bit boring and there isn’t a forest.

Image #4: Moon landscape

I’ve seen some surreal images created from AI and wanted to see what “moon landscape” would turn out like. It looks a bit like a futuristic sports arena set in a nightmare with a giant full moon floating in the middle.

Honestly, I like this one. But again, I’m not hanging it on my wall.

Image #5: Sheep in a meadow

What possessed me to do another animal image? This was another mistake. These are sort of sheep shapes, but they just aren’t right. The ones in the front of the picture look a bit like tortoise-shaped sheep. The cheerful flowers, blue sky, and green grass juxtaposed with the odd sheep shapes make this an uncomfortable image to witness.

Image #6: Cityscape storm

I like this one. It looks like some great city is being swallowed up by the ocean in a final global warming event. Very dystopian and kind of pretty.

Image #7: Selfie at the end of the world

I decided to go all out for my final image before my daily credits ran out. I must admit, I did not think of this prompt on my own, but saw an article in the NY Post about a viral TikTok of AI-generated images. The prompt was “selfie of the end of the world.” The images in the article were horrifying. And here is what mine looks like. Pretty scary, right?

Though the skull face is frightening, the long, dark hair, reminiscent of the girl from The Ring is the stuff that will give me nightmares from this one.

Who owns AI-generated images?

You may be wondering if these amazing creations are copyright free.

According to the Smithsonian Magazine, the U.S. Copyright Office has recently rejected copyright requests for AI-generated artwork because the image, “lacks the human authorship necessary to support a copyright claim.” I don’t know what this means for other countries, but the Smithsonian article suggests some countries may be more willing to put copyrights on computer-generated works.

According to the NightCafe program that I’ve used here, “Your creations belong to you, and you can do whatever you like with them (provided you are — or have permission from — the copyright owner of any input images, and subject to Copyright laws in your jurisdiction).”

So there you go. For now, unless something changes with the copyright laws, you should be free to use your AI creations in blog posts and articles. That is unless you’ve used an input image that you don’t have permission to use.

Be sure to check the copyright rules of whichever AI-generated program you are using, and be prepared for this to change someday if the law changes. It’s all very new still and things are likely to shift at some point.

I don’t think I’ll be ditching Unsplash to use AI-Generated images every day, but this was a lot of fun. That is, as long as I don’t have nightmares tonight about monster pug dogs lolling about on beaches.

The author assumes responsibility for the copyright of the images in this post.

To unlock unlimited stories, consider becoming a Medium member for $5 per month. If you sign up with my referral link, I’ll receive a small commission at no extra cost to you.