swedish artist simon Stålenhag is known for haunting paintings that blend natural landscapes with the grotesque futurism of giant robots, mysterious industrial machines and alien creatures. Earlier this week, Stålenhag appeared to experience some of his dystopian fears when he discovered artificial intelligence was being used to mimic his style.
The AI-mimicking behavior was done by Andres Guadamuz, a reader in intellectual property law at the University of Sussex, UK, who has been researching legal issues surrounding AI-generated art. He uses a service called Midjourney to create ghost-style images similar to Stålenhag and post them to Twitter.
Guadamuz said he created the images to highlight the legal and ethical issues that algorithms that generate art could raise. Midjourney is just one of many AI programs capable of producing artwork on demand based on textual cues, using machine learning algorithms to digest millions of labeled images from the web or public datasets. After that training, they could imagine almost any combination of objects and scenes, and could reproduce the styles of individual artists with astonishing accuracy.
Guadamuz said he chose Stålenhag for the experiment because the artist has criticized AI-generated art in the past and may object. In a blog post after the incident, Guadamuz argued that lawsuits alleging infringement are unlikely to succeed because while a piece of art may be copyrighted, an artistic style cannot.
Stålenhag disapproved of the stunt.in seconds series of tweets This week, he said that while borrowing from other artists is “a cornerstone of living arts culture,” he dislikes AI art because “it shows that this derived, generated goo is our new The tech lords want to feed us what is in their vision for the future.”
Stålenhag did not respond to a request for comment.guadamuz public apology to Stålenhag and said he deleted the tweet containing the derived image. Guadamuz also said he received angry messages, including death threats, from some Twitter users who disapproved of his stunt. What started out as a thought-provoking experiment was misinterpreted as an attack, he said. “I was boring and suave academic during the day, but by night I became a supervillain who ruined an artist’s livelihood…or whatever,” jokes Guadamuz.
Algorithms have been used to generate art for decades, but a new era of artificial intelligence art began in January 2021, when artificial intelligence development company OpenAI announced DALL-E, a new technology that leverages machine learning from a string of A program for generating simple images from text.
In April, the company released DALL-E 2, which generates photos, illustrations and paintings that appear to be made by human artists. In July, OpenAI announced that DALL-E would be available to anyone, saying the images could be used for commercial purposes.
OpenAI uses keyword filters and tools that can spot certain types of images that might be considered offensive to limit what users can do with the service. Others have built similar tools – such as Midjourney, which Guadamuz used to imitate Stålenhag – and they may differ in the rules for proper use.