Generative AI Helps Artists Expand the Boundaries of Their Imagination

In 2023, there are two types of creative people. Either you have some level of contempt for generative AI, or you are excited about its potential. It is rare these days to find a writer, painter, musician, filmmaker, or any other type of artist without an opinion one way or the other.

The term “generative AI” has drawn plenty of headlines lately for its use in ChatGPT, the chatbot launched last November with the uncanny ability to write in a humanlike voice. But ChatGPT is just one of many generative AI tools that writers and artists from across the creative spectrum are using to augment or, in some cases, completely upend the creative process. While many artists are busy wringing their hands about generative AI stealing their jobs and/or their intellectual property, others are busy finding ways to use it to create something wholly new and exciting.

Ross Goodwin certainly falls into the latter category. Goodwin, a former ghostwriter for President Obama and a creative technologist at Google, used generative AI to produce a book called 1 the Road. It’s an homage to Jack Kerouac’s iconic road trip novel On the Road, even following the New York to New Orleans route that Kerouac made famous, except that it was written entirely by AI. Goodwin built the AI using a neural network that he trained on hundreds of books he felt matched the style he was going for (although not the actual On the Road, since he felt like that would be cheating in some way), as well as Foursquare location data relevant to the journey.

“I went from ghostwriting to artificial intelligence, and started with learning to code,” Goodwin explains in a short film made about the novel. “I was able to put books that I liked into this machine that could suddenly learn to write like my favorite authors, and it was intoxicating.”

Goodwin’s idea was to write a book using the car as an instrument. The algorithm was fed data from the car’s GPS system, a microphone, his computer’s internal clock, and a camera sitting atop his Cadillac to produce AI prose describing the adventure. The output was something that The Atlantic writer Brian Merchant described as “the Electric Kool-Aid Acid Test meets Google Street View, narrated by Siri.”

Goodwin sees AI as a tool to help people expand their limits, creatively speaking.

“With AI, I’m most excited about the idea that I could have this companion that knows me better than anyone and that I can create things with, that is custom-tailored to my needs, that will do exactly what I need it to do to help me produce work that’s beyond my native capacity right now. For me, AI is really about giving people tools to reach beyond their native capacities.”

Shedding new light on the meaning of life

The notion of creativity refers to our ability as humans to manifest something real from its beginnings as an idea in our imagination. The output of our creativity, the artwork itself, is therefore a reflection of mankind and what it means to be human. But what happens when we introduce artificial intelligence into the creative process? What does an intelligence that is not our own but is based on our intelligence tell us about the meaning of life?

“We wish to demonstrate that algorithms help us complete our understanding of how we function as humans and push us to outsmart our current level of creativity,” said Pierre Fautrel, Hugo Caselles-Dupré, and Gauthier Vernier—three artists known collectively as Obvious. The group came to fame in 2018 when its AI portrait “Edmond de Belamy” sold for $432,000 at Christie’s auction house in New York City. It remains the highest-priced AI artwork to this day.

Writing in their online manifesto, the trio says they see algorithms as “a fascinating tool to dig into and better understand the different forces at stake in the process of creating something new, unique, and innovative.”

For “Edmond de Belamy,” Obvious wanted to make a series of portraits of a fictional family whose family tree would serve as a dual commentary on art history as seen through the eyes of artificial intelligence, as well as on the impact that AI will have on the future of society. The group first compiled a database of 15,000 portraits painted by humans over the years. Next, they built an algorithm to generate new portraits based on an analysis of what makes each piece unique. This part of the process, the group claims, required a long series of trials and errors to train the algorithm to produce results they could all be happy with. After reviewing a large number of images produced by the AI in a variety of styles, the artists selected the pieces they felt best served their message.

Obvious believes we are entering a new era of creativity for artists. “We started as three humans, limited by our creativity and our biased vision of the world surrounding us,” their manifesto explains. “We expanded our mind using algorithms by developing a tool providing us with an algorithmic view of our world.”

Creativity on steroids

Recent advancements in generative AI enable computers to do much of the heavy lifting that the artists of Obvious had to perform themselves while creating “Edmond de Belamy” back in 2018. Tools such as Stable Diffusion, Midjourney, DALL-E 2, and others have been trained on enormous datasets of images and artwork, and their algorithms are surprisingly sophisticated, producing output that is practically indistinguishable from human artwork.

Now, artists can get completely novel images in a matter of seconds based on any prompt they can dream up, enabling them to iterate exponentially faster than ever before.

“AI augments my creative process by allowing me to distill and recombine textures, and that's something that would otherwise take me months to draw by hand,” Portuguese artist Sofia Crespo said in a TED Talk last November. Crespo, whose work focuses on how humans can express themselves creatively through the use of artificial intelligence, used neural network technologies to produce “Neural Zoo,” a series of striking images that she describes as “speculative nature.” The images offer a surreal look at fictionalized plants, animals, and otherworldly organisms that could only exist in an AI-inspired, bizarro universe.

Crespo is one of many current artists who believe that generative AI can help humans expand their artistic capabilities and ultimately come up with entirely new styles of art. “Knowing that there’s a boundary to our imagination doesn’t have to feel limiting,” she offers. “On the contrary, it can help motivate us to expand that boundary further and to seek out colors and things we haven’t yet seen—and perhaps enrich our imagination as a result.”

Re-training our brains

Generative AI is being used to create new content in the music industry as well. While the automatic generation of music dates back more than half a century, recent advances in generative AI have brought a whole new level of sound quality, composition, and vocal effects that were impossible just a few years ago. AI tools such as Jukebox, Boomy, AIVA, Soundraw and others, are among the top platforms available today to help musicians and non-musicians alike create new songs with just a few prompts. Even Google announced earlier this year it has built an AI system, dubbed MusicLM, that can make music in any genre based on a text-based description, although the company says it has no plans to release the service just yet.

One of the main benefits of using generative AI, according to some musicians, is to push them out of their comfort zone. Musicians tend to rely on physical expressions that have been ingrained into their minds and bodies over many hours of repetition —  certain riffs or chord changes that become second nature and appear over and over throughout their music with only slight variations. But AI might come up with something novel that forces them to come at their music in a completely different way.

When the dance-pop band Yacht made its Chain Tripping album, it relied on artificial intelligence to help write the lyrics, compose the music, and even create the artwork, videos, and typography used to promote the album. The band trained AI algorithms on its own back catalog of 82 songs to come up with several hours’ worth of new melodies and lyrics. They then stitched together various fragments of music to create coherent songs that sounded both familiar and exotic. The album resulted in a Grammy nomination, the band’s first, for best immersive audio album.

“I think musicians often underestimate how much the way we play is based on our physical experiences and habits,” Yacht member Claire Evans told TIME. “AI forced us to come up against patterns that have no relationship to comfort. It gave us the skills to break out of our own habits.”

A new definition of creativity?

Generative AI is still in its nascent stages, and for now, its use in the creative process remains very much an experiment. But on top of some of the traditional creative realms described here, artists in fields as diverse as filmmaking, interior design, and even coding have also been using the technology to brainstorm new ideas, help get unstuck, and produce something new and unexpected — something ostensibly better than they could’ve come up with on their own.

And while there are plenty of ethical, legal, and practical questions that must be answered as its use becomes more mainstream, it is clear from these early experiments that generative AI has the potential to not only enhance the creative process but to redefine the very idea of creativity as we know it.