AI is expanding the world of illustration. Here’s how artists are competing | Tech Rasta


This is a sight that was unimaginable a few years ago.

Earlier this month, a popular Korean-language artist who goes by @ato1004fd on Twitch live-streamed an 11-hour sketch session, watched by their 22,000 followers as they created a picture of a popular character from a video game. The effect of ginseng.

But by the time @ato1004fd finished the digital painting, a rogue viewer had already grabbed an image of the work-in-progress from the stream, used AI to “finish” it, and posted their own version on social media. Accusing @ato1004fd of being a copycat.

“Bro, when you ask your fans to cry about stealing art, [be] Reasonable,” Forger wrote. “So you took the AI ​​picture as a reference, but at least admit it.”

Backfired Twitter There the scammer accused the users of gaslighting the original artist. Eventually, the faker deleted their account.

Recent developments in machine-learning programs have transformed AI into an incredible artistic tool capable of outperforming human artists and cutting costs, creating an earthquake in creative circles. Concerns are high among graphic artists and commercial illustrators whose livelihoods are tied to the ability to alter content to clients’ specifications.

“It’s the Wild West right now,” said Liz DeFiore, illustrator and president of the Graphic Artists Guild. People are mainly worried about their jobs.

DiFiore cited information from price guides his firm published last year, which listed the median salary for a painter at $50,000, about $10,000 less than what the Bureau of Labor Statistics estimates as the median annual wage for artists.

“Wages have been going down for years,” she explained. “And the biggest reason illustrators see the value of their work drop is probably infringement.”

The anime art fraudster used an image generator called NovelAI, a monthly subscription service that promises algorithm-assisted authoring and storytelling. The program includes an “image to image” service that allows users to upload images that can be edited via text prompts given to the machine. “AI must learn how a person creates—just like a person does,” the company explained in a recent blogpost with the bold “Please use our tool responsibly.”

AI image generators learn to make artworks based on large datasets with billions of images scraped from the Internet. Many artists are unaware that their paintings and prints exist in these vast online archives; Companies believe that they do not need to disclose the contents of their vast archive as the information is largely considered public domain.

Greg Rutkowski, Dragon's Breath (2016)

Greg Rutkowski, Dragon’s Breath (2016) Courtesy of the artist.

In September, Poland-based artist Greg Rutkowski, a prominent name among fantasy art and gaming fans, criticized the system after learning that it had become one of the most popular prompts for image-making algorithms. His name was used more than 93,000 times and people began mistaking the computer’s output for his own work. A platform, Disco Diffusion, suggested his name as a prompt.

“AI should exclude living artists from its database,” Rutkowski said at the time.

Popular programs like DALL-E, Midjourney and Stable Diffusion already have policies in place that prevent users from using their products in certain ways, such as assigning images of celebrities and politicians.

Matt Dryhurst and Holly Herndon

Picture of Matt Dryhurst and Holly Herndon. Image courtesy of the artists.

As the powerful and wealthy have a way to escape online image scrapers, many artists have begun asking companies to remove their work from AI training models. Artists Matt Dryhurst and Holly Herndon recently launched Spawning, a tool intended to allow people to set permissions on how machines can use their style and likenesses.

“After years of beating our drum about the importance of training data, there is now an appetite for discussion,” Dryhurst told Artnet News.

“I understand why artists involved in commercial work are concerned,” he added. “If there is a fear that these machines will create satisfactory art (beautiful paintings, sculptures with modern accents), then yes, that fear is valid. I believe that kind of art will become automated in the next couple of years.

Spawning involves a search function that searches through a dataset called LAION, a 150-terabyte dataset that most AI image generators use for training. A quick search of living artists returns thousands of images in the archive of works by Peter Doig, Damien Hirst, Jenny Holzer, Alex Katz, Jeff Koons, Kerry James Marshall, Cindy Sherman and many more.

Screenshot of the results for "Cindy Sherman" From spawning

A screenshot of the “Cindy Sherman” results from the Matt Dryhurst and Holly Herndon website spawning.

The rapid adoption of artificial intelligence has surprised many major galleries, which typically have strict controls over who can use images by their artists. Representatives for Pace Gallery, Gagosian and Spruth Magers declined to comment on their policies regarding AI image generation.

Legal experts expect artists, illustrators and galleries looking to challenge AI image generators to face an uphill battle in court. According to Megan Noah, an attorney and co-chair of Pryor Cashman’s Art Law Group, there isn’t much precedent when it comes to copyright infringement and artificial intelligence.

“The field is very dynamic and evolving,” she said, explaining that plaintiffs need to prove substantial similarity between the original artwork and the replica beyond a familiar style. “An artist has no copyright protection over style,” Noah explained.

Lacking legal recourse, commercial artists have turned to commercial organizations for help in lobbying the tech companies behind AI image generators. DiFiore said her guild has experienced an uptick in membership over the past few weeks because of the concerns. “I spoke to at least three new members who said they joined the guild because we were talking about AI”

Despite the concerns, DiFiore still hopes the technology can be used to improve the lives of the graphic artists she represents. “AI can take some of the drudgery out of our jobs, like resizing images, creating ad designs, and looking at compositions from different angles,” she said. “These are things that often take time away from the parts of the job that we enjoy. Any professional would appreciate the ability to get to their jobs—while maintaining their wages.

Follow Artnet News on Facebook:

Want to advance in the arts? Subscribe to our newsletter to get breaking news, eye-opening interviews and key takeaways that move the conversation forward.


Source link