Artificial Intelligence art is disposable, and it’s ruining art

I paid an artificial intelligence (AI) art generator $12 to create 424 portraits of myself spanning 53 art styles and historical periods because I watched strangers on TikTok do the same.

The number of images I'd received intrigued me the most, the possibility of seeing myself in moments across thousands of years I'd never lived was really exciting. Paying a one-time fee for nearly 500 images sent to me in minutes was just tragically too convenient to ignore.

After I poured over each image — sharing the most hilarious with my friends, of course — I closed the website and moved on. Did I frame any of them? No. Did I prepare a scrapbook? No. The momentary joy sustained by the instant gratification wore off far quicker than it took to input my payment information.

The danger of this transaction was not the amount of money wasted — what's most concerning is that each image was generated by stealing the work of artists instead of commissioning a real, living artist.

I bought 424 products of algorithms, but I did not purchase art.

A picture is worth a thousand words because there is craftsmanship, flaws, heart and perspective involved in creating art. What story does an AI-generated work even tell? The lack of humanity in AI-generated art compromises everything about art that is treasured.

AI is more closely related to a calculator than any sentient being with thoughts, emotions and experiences to draw upon. No piece of art will ever be perfect because perfection is not art’s purpose. Art allows for individuals to express their thoughts, desires, memories and, especially, pain in a physical medium. AI generators could use the most pristine and well-crafted pieces of art ever created, yet still any product created will have no story to tell. Any art created entirely with the use of stolen art will never be art because these generators allow the lowest amount of human involvement.

As they stand currently, AI art generators are not an instrument to create something new, but a vile attempt to replace the artist entirely.

Painter and professor Harold Cohen created AARON, an AI to autonomously generate images in 1973. However, the release of Stable Diffusion has granted everyone the power to create hundreds of their own images within seconds. Released in August 2022, the deep learning model marked the first open-source text-to-image model that was free to use. Stable Diffusion draws from LAION-5B, an art database containing 5.85 billion image-text pairs taken from the internet.

The algorithms appear harmless — seemingly no different than applying a Snapchat filter, allowing users to travel through time, create magic avatars and advanced beautifying filters.

However, the product of these algorithms have already affected the livelihood of real artists. At last year’s Colorado State Fair’s annual art contest, an AI-generated picture won first prize with the use of Midjourney, a program that creates images from text descriptions with the help of Stable Diffusion.

Beyond compromising the sanctity of art, the public release of Stable Diffusion has inspired a rush of interest in AI-generated images, which has instilled genuine fear in real artists.

Clarabelle Sullivan, sophomore studio arts major, is passionate about painting portraits and fears this rise in visual plagiarism is degrading the appreciation of real artwork. “If I saw my art — or something that looks like my style [used to create AI-generated work] — I would be pretty shocked, and it is worrying, especially ... being an artist. It's … a business, it's a job; people make a living off of it,” she said.

Sullivan’s fear has already become reality — Artist Deb JJ Lee discovered they were the victim of art theft, enabled by capitalize stable diffusion after being sent AI-generated artwork created using their own art. “This is f------ sad. I don’t know what to do,” said Lee on Twitter.

Finding inspiration in another artist’s work is already a fine line to walk, but to explicitly pull from a non-consenting artist to create new work without permission is art theft.

Stable Diffusion confirmed that artists are not allowed to remove their art from the database. “There was no opt-in or opt-out for the LAION-5b model data. It is intended to be a general representation of the language-image connection of the Internet,” read their frequently asked questions.

Maintaining a language-image connection to the internet is a pathetic excuse. If a living artist stole another artist’s art, there are real legal consequences. At the very least, if a living human being did steal art, they possess the human ability to fix their mistake by crediting the artist.

Replacing human labor with AI has already permeated several industries and will continue to grow in profit. AI isn’t going anywhere, but there is still a chance to change how it’s used to produce art.

Professor of multimedia arts Dmitry Kemell envisions a future in which AI can be used to create art, not as a generator. “We're living in a world of exponential growth and exponential numbers in everything — too many options. [AI] can sort of help to [assist] us, and then we can start to sort of function as art directors,” said Kemell.

The path ahead is clear, it is imperative that AI is solely used as a tool in which artists are in control instead of using generators that steal art without consent only to pass it off as original.

AI generators may be fun, but if used irresponsibly in replacement of true art, they will continue to make art disposable. Support real artists instead of a hasty and uninspired amalgamation of their work.

This is the opinion of Kylie Clifton, a sophomore journalism major from Dublin, Calif. Email comments to editor@theloyolan.com. Follow and tweet comments to @LALoyolan on Twitter, and like the Loyolan on Facebook.