Pin It
320408535_845388896686719_8768236600223072257_n

The AI designer creating fashion grails from iconic runways

Field Skjellerup’s hyper-realistic designs look like they’ve been downloaded from a cybernetic mind bank

Though their eyes have a holographic quality and their flesh appears to have been oxygen-starved, there’s an uncanny magnetism to the models that Field Skjellerup works with. The fashion archivist has been experimenting with AI for the best part of two years, feeding images of vintage Japanese fashions through a GAN model – a recent innovation in machine learning – to eerie and convincing effect. He’s created sculptural puffers with slashed waistbands, metallic concertina skirts, and avant-garde parkas festooned in artfully-arranged piles of rubbish. Prompted by images of Junya Watanabe AW04, Hussein Chalayan’s AW00, and Kosuke Tsumura’s Final Home project, the images look as though they’ve been pilfered from some kind of cybernetic mind bank. 

“I’ve spent the majority of my adult life working dead-end jobs for minimum wage and I have little to no relationship with any educational institution,” he says. “But this is such a powerful tool. I’ve managed to create the blueprint for the most hyped pair of sneakers in the world, and I think that’s really saying something.” Skjellerup’s referring to a series of Nike shoes that he showcased on Instagram last week, which look almost exactly like the kind of thing Simone Rocha would design, surfaced in laser-cut mesh, rubberised petals, and ribboned laces. Far from the haunted DALL-E renderings that have been popularised online – all scorched edges and Francis Bacon wails – Skjellerup’s creations manage to look real. “People think they’re physical products and have asked me how to purchase them. I’m just waiting on the call from Nike!”

As exciting as this may be for STEM enthusiasts, the advancements made in AI fashion often give way to real-life anxieties. If a machine can learn the handwriting of a brand and outpace its atelier, what’s to say that designers won’t be rendered obsolete? “I feel like people’s fears are warranted but the idea that artists will be eradicated is a little far-fetched,” Skjellerup says. “There will be job losses but I think new movements will arise from automation, people will put an onus on ‘human-made’ products. Am I setting an example for this kind of thing? I’m not sure I have an answer for that yet.” If creativity was once understood to be a uniquely human accomplishment, then AI has also inherited its prejudices – regurgitating images of rakish models with the spindly legs of a vaudevillian puppet. The algorithm can make our lives easier, but it’s often as flawed as we are.

Ultimately, humans choose how these systems are made and what data they are exposed to – and the consequences can often be severe. In January 2020, Robert Williams, a Black man in Detroit, was arrested for a crime he did not commit because of an incorrect facial-recognition match. AI (much like the cameras they have been trained by) are rubbish at recognising people with darker skin tones, which is a standard that dates back to when film cameras were optimised to capture the faces of light-skinned people. “I’d admit this has been the trend with the majority of my published imagery to date,” Skjellerup says. “Given the racial bias in these pre-trained models, it’s easy to fall back on whiteness as the default. My hope is that the companies creating them will change but I’d personally like to make improvements in producing imagery that includes all kinds of people.”

For Skjellerup, who also runs an impressive sales platform of archival Comme des Garçons, Issey Miyake, and Yoshiki Hishnuma (amongst hundreds of other grails) AI is first and foremost a research tool. Albeit somewhat fictional, his work has the potential to breathe new meaning into some of fashion’s most seminal collections – locked into a constant process of sampling and recontextualising. It means that even the most legendary designers will be able to live forever – their body of work absorbed and reproduced by the algorithm. The development of AI tools will mark the biggest change in all mediums of art for the next 100 years,” he says. “Integration will be fraught with ethical and legal discussions and radical new ways of art making. Not only will people use these systems to play into pre-existing processes but new modes of expression will also be created.”