Artificial intelligence (AI) image editing apps are skyrocketing in popularity, and artists’ copyrights are shrinking along with them.

You’ve probably seen the impressive and strange products of Dall-E, Midjourney, and Stable Diffusion, so-called AI image makers that generate images from a soup of existing works scavenged from the internet. Until now, AI art has been a curiosity, a tool for nerdy artists, and a fun way to waste some time. But now apps are popping up that charge money to generate images from those copyrighted sources without permission. Things are about to get ugly.

“The problem is that AI art isn’t really ‘art.’ It’s not created by a human’s imagination. Instead, it’s assembled from countless pre-existing works of art. It’s a more complicated paint-by-numbers scheme. The more consumers use these apps and support AI instead of real, actual artists, the closer we come to the death of real art. It’s going to happen piecemeal, and it’s going to be impossible to stop because proving your art is stolen in order to generate new art is going to be nearly impossible,” artist and graphic designer Amy Weiher told Lifewire via email.

To use one of these image tools, you type a text prompt describing the image you want, and it creates it for you. It can do this because it’s been trained on millions of existing images from the Internet. If you ask for an image of a doll from a 70s children’s TV show, it doesn’t know what a doll is, or the 70s, or even a TV show. But by scraping all that human-made work and noting the text and descriptions around it, it can serve up something similar.