Help! The Robots Are Coming

DALL•E 2 Generated Image with Input: Renaissance Painting of AI Killing Artists (Courtesy of Matthew Peterson)

The opinions reflected in this OpEd are those of the author and do not necessarily reflect the opinions of staff, faculty and students of The King's College.

 

Art generated by AI has been an ongoing topic of discussion for several years, with content generators dominating social media and becoming the subject of many philosophical debates about art. 

Despite the growing excitement surrounding AI-generated images, there are concerns about their authenticity and ethical implications. Many artists argue that AI-generated art is unethical as it relies on being fed pre-existing art and recreating it until it is indistinguishable from the original. The fundamental question remains: can AI create art, and what does it mean for the future of the art industry?

Believe it or not, the two paragraphs you just read were written by ChatGPT. All I did was type “Write an introduction in 60 words or less for the following article…,” paste the rest of this piece into OpenAI’s famed Generated Pretrained Transformer (GPT) and voila – no need to do my job.

But wait: is that really what I want?

Do I want a computer to be able to produce something that I’m paying thousands of dollars to learn how to produce?

 Well, no, not really.

Though people have been experimenting with them in their current rendition since 2017, artificial intelligence (AI) content generators have dominated Twitter threads and Reddit pages in the past year. This is largely thanks to Dall•E Mini, an AI art generator created in large part by Boris Dayma. 

Dall•E 2 Generated Image with Prompt: Political Cartoon of Margaret Thatcher at the bottom of a deep well (Courtesy of Matthew Peterson)

Dayma’s project was activated in July 2021. His goal was to create a public version of the privately distributed (at the time) Dall•E 2. According to him, his “most creative prompt was ‘the Eiffel Tower on the moon’”. So, in 2022, it probably came as a shock to him when people started producing stuff like, oh I don’t know, ”Fisher Price guillotine.”

The immediacy with which people could create absurd images was revolutionary for a meme-obsessed culture. People were curious, and Dayma’s program was ripe for exploration.

Since Dall•E 2 actually did become (pseudo) free to the public in September 2022, AI art experimentation became even more rampant. Dall•E 2 is the real deal, and people have extensive lists of prompts to feed the new machine. 

The images are often referred to as “art.” There is little else to call them in a cultural understanding which classifies nearly anything aesthetically pleasing and stylized as art. This stirred many other conversations about art. 

These recent conversations are actually very old conversations. What is art? Who can produce it? Where is the line between influence and stealing? These have always been complicated questions, and it seems that AI-generated images may have further complicated them.

Not only are these questions about art not new, they aren’t even new with AI art. Back in 2018, “Portrait de Edmond de Belamy” became the first AI image to be sold at auction. It sold for $432,000. This was, perhaps, the moment AI began to seriously play a part in philosophical art conversations.

Edmond de Belamy (Courtesy of Wikimedia Commons)

In another triumph for AI, Jason Allen, founder of a board game company called Incarnate Games, entered his piece Théâtre d’Opéra Spatial into the Digital Arts category at the Colorado State Fair. The competition took place in September 2022 and the judges awarded Allen’s Midjourney-generated piece first place. 

Many art enthusiasts were outraged by what they considered to be “shameful” and “disgusting.” Perhaps, most significantly, one Twitter user commented “DO YOU EVEN KNOW THAT THOSE AI'S use pictures & art that belong to other people in order to stitch the "art" given in the prompt!?

This argument is one of the fundamental problems that many artists have with AI-generated images. Most Artificial Intelligence image-generating programs function by being fed pieces of art, and then they are  taught to recreate it until what they create is indistinguishable from the original art they were fed. Bits and pieces of many artworks are compiled to create any “original” work made by AI. Many feel that this could be unethical toward the original artists.

Dall•E 2 Generated Image with Prompt: Queen Elizabeth I enjoying pie and cigars (Courtesy of Matthew Peterson)

A frequent counter-argument to this concern is that art has always been a kind of thievery. Steve Jobs once famously said that “good artists copy, great artists steal.” It is no secret that art is always inspired by something, whether that source of inspiration is as broad as the beauty of red-rock canyons or as specific as Vienna by Billy Joel. 

Art comes from an exterior perception. But that doesn’t seem to make the a-computer-is-plagiarizing-my-artwork pill any easier to swallow.

One artist named Deb JJ Lee with the Twitter handle @jdebbiel posted a screenshot of messages she had received. The messages regarded someone who believed a programmer had created a “custom model” which appeared to be directly mimicking @jdebbiel’s unique style. In response, @jdebbiel simply said a defeated “this….this is f***ing sad. I don’t know what to do.” 

An AI can be trained on any art style, so it is entirely feasible that the artist’s style had been directly ripped off. It is worth mentioning that it was never confirmed whether or not this was the case. Regardless, many are understandably uncomfortable with the notion that AI can instantly replicate something they have spent years of blood, sweat and tears on.

Dall•E 2 Generated Image with Prompt: Picasso Painting of Dietrich Bonhoeffer singing his heart out at Karaoke night (Courtesy of Matthew Peterson

Dr. Stephen Salyers, or, as his students know him, “Doc,” has been a professor at The King’s College since 2000 and specializes in Media, Culture and the Arts. Doc’s office is covered from floor to ceiling in pop culture paraphernalia — most of it Disney — as well as books and artwork. The room oozes creativity and seems to hum with curiosity. It is the personal space of one who greatly appreciates wonder.

Doc’s attitude is a cautious-yet-interested one. “It really intrigues me,” he said. ”But I do think, as it gets more sophisticated…what’s next?”

“I think it is capable of creating something aesthetically pleasing to people,” Doc said when asked if AI can create art. “I think it’s capable of—at least at some point—creating something revealing about humanity.” 

He believes that the emotion and experience behind artwork are often what make it powerful. “When someone creates something artistic, there’s motivation behind it,” he said. “What does it look like when AI creates something but it hasn’t been created through the experience of that thing? Or will we someday create something that has sympathy for humans? Then we get into really weird and wonderful and scary territory.”

Doc’s simplest concerns are similar to those of the angry tweeters and most internet artists.

“It’s becoming clear to us that AI is drawing from other sources that it has scrubbed through the internet—things that people have posted,” he said.

He further explained his frustration with where this could lead by saying, “I have less of a problem if a generic artist or pseudo-artist is sitting at a computer and typing in ‘make me a composite Disney character—off of these five characters make a new one’….than I do Google having an app where Google then makes money… but they’ve scrubbed the internet… pulling out pieces of somebody else’s work. So I guess part of it is attribution and, as Scripture would say, making sure a worker is ‘worth his keep’.”

Dall•E 2 Generated Image with Prompt: Child's Drawing of the members of the House of Reagan smoking an egregious amount of cigars (Courtesy of Matthew Peterson)

The worry about making sure artists are recognized and paid for their work is not Doc’s biggest concern. 

“As a professor, the main thing I’m worried about is: I don’t want people to not create because some other entity is creating,” he said. “I think we lose our humanity when we stop creating.” 

Doc then tied this into theology by expressing the fact that in the Bible “the first thing that has been revealed to us is that [God] is a creator. I think there’s a certain beauty and wonder in that.”

Doc’s thoughts evolved beyond AI and art and began to touch on the digital world in general. 

“It seems like we’re shifting from having an interpersonal experience with one another to a digital experience with one another,” he said. “We’ve gone from talking to texting, we’re moving from hanging out with each other to hanging out online. And that can be really fun, but there is something to be said about being with other humans and creating with other humans and talking with other humans.” 

“The thing that worries me the most is that we just give away our humanity to digital entities,” Doc said. ‘We’re just slowly, incrementally giving these things away. So, if we get to a place where an entity is creating all this stuff for us, it’s one more thing we’re giving away.”

Whether it all scares or excites you, I hope this ChatGPT-generated haiku will encourage you:

Worry not, my friend

Computers will not take over

Human hearts prevail

Then again, maybe that’s just what they want us to believe.

Dall•E 2 Generated Image with Prompt: Oil Painting of Winston Churchill playing poker in Central Park (Courtesy of Matthew Peterson)