Local art exhibit features artificial intelligence

Artistic images generated from artificial intelligence (AI) are prominently featured at an art showcase gallery show in CSPS Hall in Cedar Rapids.  The gallery was created by Chris Sherman, a photographer and digital artist from Marion, who partnered with CSPS Hall to display 190 works starting Nov. 14 and ending Dec. 18. Free and open to the public, his pieces include drone and traditional camera photography, as well as the gallery of AI-assisted art. The images are just one example of a sudden advancement in AI technology — seen elsewhere through chatbots — that is captivating the internet's attention in recent weeks. For Mr. Sherman, AI imagery tools allow him to create complex works of art within seconds and edit images into different styles on a whim. The only limitations? A person’s imagination and their ability to convey that to the machine. "It's really kind of mind blowing," said Mr. Sherman, describing the evolutionary process that takes place. "The idea that you can start with a prompt [or text description] and then add modifiers and variations to that to achieve your output." While the application of AI technology is revolutionary in scope and could stretch into countless fields, some creators believe the advancements harm artists more than help by devaluing works of art. Others argue AI is not creating art but rather curating it from other works found online. https://www.youtube.com/watch?v=TT4PGksN6aA

The CSPS exhibit

Mr. Sherman, a professional photographer since 2015 and founder of commercial drone company Flying Camera Media, primarily uses DALL-E 2, Jasper, and NightCafe for his work involving AI. "My background has always been at the intersection of creativity and technology," he said.

All you need is your imagination and a text box and these tools can create images in all types of styles — from realism to abstract to oil painting and more — making even the most novice creator an artistic professional. He then runs the creations through traditional editing tools like Adobe Photoshop or Lightroom until he finds it presentable and ready for sale. Some pieces are available for as little as $15. About 90% of the digital AI-inspired pieces at the exhibition are designed for kids and include monsters, animals, fairies, dragons and more. Some images require just minor cropping and editing, while others need more complex descriptions to get the perfect AI generation.

"If you look at the trial and error that goes into it, I probably have 3000 digital works of art that didn't make the cut," he said. In the Colorado State Fair this year, the winning entrant submitted a piece of AI art, reported The New York Times. It is just one example of AI disrupting the art world. "Technology has evolved so rapidly that you can step into different worlds and do a bunch of different things that as a photographer I couldn't realize previously," he said.

Is artificial intelligence-assisted art a net positive?

AI art could drastically change how companies operate, Mr. Sherman explained. While most AI art programs are not free, he said, he envisions a time when companies may eventually decide it's more cost-effective than retaining an artist or designer. Other content writers could use AI art to create a beautiful header image for a blog post or promotional material for marketing purposes. "It's going to put an interesting bit of pricing pressure on stock photography," he said, referring to stock photo websites that often require a subscription. But do these tools unlock expression for the masses, or does AI-assisted art devalue a conventional artist’s work?

Tools like Dall-E and Jasper are built from databases that scans billions of images on the internet. An analysis is completed, almost instantaneously, to match the user's description as best the machine can. "Certain artists say 'my art has been used as a foundation for these AI tools and I didn't give permission for that,'" he said. Proponents of AI generation say technology is always being used to create images. "Everything I do with photography has a technology component – whether it's flying a drone or post-processing through an editor," Mr. Sherman said. "There has always been some software manipulation." In the longer-term, Mr. Sherman foresees AI being capable of generating photorealistic images from different times in history, like the Great Depression or 18th century. But since these images would be fake and will be hard to differentiate from actual historical photos, it's easy to envision a time when no image is trusted and every image is scrutinized.

No limit to artificial intelligence capabilities

Mr. Sherman said in the past he has considered creating a children's book but never pursued the path. With AI art, this option becomes more readily available. A computer scientist in California had the same idea. Using ChatGPT, a language-generation chatbot that spits out answers in the form of text the same way DALL-E does for images, Ammaar Reshi created (and then edited) a children's tale, reported BuzzFeed. He then used MidJourney, another AI image tool, to make illustrations to match the story AI created for him. ChatGPT is developed by OpenAI, the same company that founded DALL-E and its successive iterations. It was founded by Sam Altman, Elon Musk and others. Its funders have committed more than $1 billion to the efforts, according to OpenAI's website. Examples online show chatbots writing essays, solving puzzles and telling jokes. Another AI image tool called Lensa is an app, powered by Stable Diffusion, that allows users to take a selfie and then witness the selfie become rendered in different artistic styles and themes. According to Morning Brew, it was one of the most downloaded apps after launching in late November.

Some critics argue that uploaded images are being used to train its neural networks that then help create more images, meaning personal data is not protected. Other participants said the images Lensa creates are highly sexualized and can struggle to work for people of all skin colors, reported BuzzFeed. Despite OpenAI working to curb prejudicial responses that occur through its companies, unfiltered bias still occurs, reports The Intercept. Whether good or bad, the genie could already be out of the bottle, said Mr. Sherman. "I think this is a natural progression," he said.