How KEZIA Barnett creates her whimsical works of art using AI

There are pictures of beds covered with rocks, of a female figure who is buried in the earth up to her neck, another whose face is only just above water. The images were created as Barnett lay in bed deeply unwell, using only her phone and occasionally her laptop.

Barnett creates artworks using artificial intelligence (AI). She inputs text prompts – keywords describing the kind of image she wants – into an image-creation tool that generates a new picture, rendered using the program’s vast datasets of existing imagery.

The result is a kind of co-creation between Barnett’s imagination and the AI’s visual interpretation of the text. Barnett iterates again and again until she is happy with the image. Or she abandons it and starts again.

She can spend hours on each image, editing, selecting and refining. The AI program will never produce the same image twice, even with identical text prompts, as the algorithm is always evolving and improving.

Barnett, 47, is an award-winning director of short films, commercials and music videos who enjoyed a successful career both here and overseas until she became seriously ill.

In 2011 she was in South Africa directing the biggest job of her career when she contracted a virus, from which she never fully recovered. A short time later, she sustained a severe concussion, which continues to affect her to this day.

She had been based in London, but too sick to work, returned to Auckland.

“For a long time nobody seemed to know what was wrong with me. I kept on trying to create and work, not knowing it was making me sicker.”

Barnett would eventually learn that her concussion was a lot worse than originally believed and was diagnosed with post-concussion syndrome. This was in addition to a diagnosis of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), a complex condition with a bewildering set of symptoms, including debilitating fatigue, cognitive dysfunction, sleep disturbance and pain – all made worse by exertion of any sort – and often triggered by a traumatic injury or virus.

Barnett has spent most of the past few years in bed.

“When I’m at my sickest I get very sound sensitive. I get light sensitive. I get a lot of body pain. Being upright feels upside down, so I have to lie flat.

It feels like my blood is lead, like I’m buried alive, like I’m underwater. Cognitively, I feel very far away.

While I still have all my cognitive faculties, I can’t do as much as I’d like due to limited energy. I am about 95% bed bound at the moment and the isolation can be extreme. I crash really easily if I push beyond my energy envelope.”

Barnett only began creating AI art in July and has since made thousands of images, with only some making the final cut. She says Web3 (a new iteration of the world wide web that hosts decentralised apps, which run on blockchain technology) has opened her up to a huge online community of interesting, like-minded people.

Many of whom are living with various disabilities and health challenges, too. “There’s a whole level of accessibility that the Web3 world offers. Because I’m no longer able to be on film sets, I’ve been looking for a way to continue creating imaginary worlds and telling stories.

I’ve been wanting to find an artistic expression of my experience, to raise awareness for the millions of people around the world that are suffering, and do it in a way that people want to look at it, and not just look away.”

Flying Bed is one of three series, along with Bubblegum Queen and Ethereum Forests, which together comprise the Fairytales Dreams collection.

“They’re all inspired by my experience with Web3 in some way.” Each of Barnett’s images is either a character, location or a scene set in an imaginary fairytale world. She invites collectors to participate in shaping the narrative. They can help name the characters, describe them and have a say in what happens, making use of the blockchain technology and the freedom of direct connection between artist and collector.

“Working with AI feels like directing in some ways, instead of directing a crew, you’re directing the AI. With both there is collaboration.”

All the works are one-of-a-kind and are for sale as NFTs (non-fungible token is a unique digital identifier recorded in a blockchain and is used to certify authenticity and ownership). Barnett’s images sell for 0.1 Ethereum, which is approximately NZ$235 (but this can fluctuate wildly).

Predictably, the arrival of AI art has caused consternation in the art world – just like the introduction of Photoshop before it, and digital photography before that. It invites the question: what makes art, art? And who gets to decide that anyway?

Some art communities and marketplaces are banning AI-generated images entirely, along with galleries who say they won’t show AI art. In August, an AI artwork won the digitally manipulated photography category at the Colorado State Fair’s fine art competition. Cue outrage. The winning artist, Jason Allen, declared to The New York Times: “Art is dead. It’s over. AI won. Humans lost.”

Closer to home, Gary Langsford of Auckland’s Gow Langsford Gallery says: “We certainly wouldn’t not take on an artist because they were using a component of AI in their practice. There’s good art that’s made with artificial intelligence and there’s really bad art.

It comes down to how the artist is using the technology and the image that they are actually making. That’s in the eye of the beholder to a large degree. If it’s us, then we are making value judgments on its quality.” Ultimately, says Langsford, it’s just another creative tool in the artist’s toolbox.

When it comes to making sense of the legal ramifications of AI art, solicitor Julius Hattingh, who specialises in intellectual property at Hudson Gavin Martin, says there’s a lot of uncertainty and ambiguity around AI art because it’s such new territory.

“The headline is, it’s complicated. The core idea is if you create something, then you should own it. But who is the creator? Is it the person who created the AI code? Is it the person who creates the work that is then fed into the AI that allows it to create resultant works? Or is it the user of the AI tool that inputs an interesting prompt to create a new piece of art?”

Hattingh says there is an important concept around derivative works at play. “Does it meet the threshold of an original piece of work, have you added enough originality to it? Is there sufficient skill and effort from the user, or is it just arbitrary and random?

Where the AI tool functions by drawing upon existing works of art, then this leads to an interesting philosophical tension between mere inspiration and copying. Overlaying all of that, he says, will be the terms of service of the individual AI platform that’s being used.

For Barnett, AI is more than just a new creative tool.

“It’s given me a way to express myself, the freedom to communicate, and brought me a sense of hope and joy. Creating is my reason to be, it’s been my drive, it’s been my meaning. Without it I feel incomplete. So to find something that enables me to create is incredible.”