Rutkowski was initially surprised, but thought it could be a good way to reach new audiences. Then he tried to look up his name to see if a piece he had worked on had been published. The online search brought up a job that had his name attached but was not his.
“It’s only been a month. And a year from now? I probably won’t be able to find my job out there because [the internet] it’s going to be flooded with AI art,” Rutkowski says. “That’s worrisome.”
Stability.AI, the company that created Stable Diffusion, trained the model on the LAION-5B dataset, which was compiled by the German non-profit organization LAION. LAION gathered the data set and reduced it by filtering out watermarked images and non-aesthetic ones, such as logo images, says Andy Baio, a technologist and writer who downloaded and analyzed some of the Stable Diffusion data. Baio analyzed 12 million of the 600 million images used to train the model and found that a large portion of them came from third-party websites like Pinterest and art shopping sites like Fine Art America.
Many of Rutkowski’s artworks have been pulled from ArtStation, a website where many artists upload their portfolios online. Its popularity as an AI indicator stems from several reasons.
First of all, its fantastical and ethereal style looks very beautiful. He is also prolific, and many of his illustrations are available online in high enough quality that there are plenty of examples to choose from. An early text-to-image generator called Disco Diffusion was offered by Rutkowski as an example indication.
Rutkowski has also added alt text in English when he uploads his work online. These image descriptions are useful for visually impaired people using screen reader software and also help search engines classify images. This also makes them easy to scrape, and the AI model knows which images are relevant to the prompts.