Years have been spent by Erin Hanson refining the brilliant color scheme and bold brushstrokes that characterize her well-known oil paintings. However, I demonstrated to Hanson my attempts to imitate her style with only a few keystrokes during a recent interview with her. I had entered a series of prompts into Stable Diffusion, a well-known and widely used open-source AI image-generating program, to produce pictures in the manner of some of her paintings of California poppies on an ocean cliff and a field of lupin.
She commented on one of my attempts via Zoom, “That one with the purple flowers and the sunset absolutely looks like one of my paintings, you know?” I then customized a second specific request using Hanson’s advice: “Oil painting of crystal light, in the manner of Erin Hanson, light and shadows, backlit trees, strong outlines, stained glass, modern impressionist, award-winning, trending on ArtStation, vibrant, high-definition, high-resolution.” I gave Stable Diffusion the prompt, and it generated three photos quickly.
She exclaimed, “Oh, wow,” as we went at the pictures, pointing out how the trees in one image resembled those in her 2021 painting “Crystalline Maples.” She quickly added, “I would put that on my wall.”
The data set used to train Stable Diffusion, which was released in August by London-based Stability AI, included the work of numerous professional artists, including Hanson, who is based in McMinnville, Oregon.
Text-to-image AI systems, which were formerly exclusive to a small number of tech insiders, are growing in popularity and strength. These solutions come from two different companies: DALL-E, which has raised $1 billion to date, and Stable Diffusion, which recently received more than $100 million in funding.
These programs may produce a variety of images with only a few words, including some that are obviously suggestive of the works of many, many artists. They normally offer a few free credits before charging. By using terms like “in the style of” or “by” combined with a specific name, users can invoke those artists. These tools are currently used for a variety of purposes, from business uses to just fun.
Text-to-image AI systems have attracted millions of users in a short period of time, and they are already being used to produce experimental films, magazine covers, and graphics to accompany news articles. A recent painting competition at the Colorado State Fair was won by an artwork created by an AI system named Mid journey, which infuriated artists.