In the ever-evolving landscape of fashion, designers and brands are increasingly weaving the threads of technology into their collections.
The intriguing interplay between art and Artificial Intelligence (AI) has sparked the curiosity of a few designers and in the previous months we have seen Coperni's A/W23 collection featuring human paintings of a robot and a lamb generated with AI text-to-image software DALL-E while Marc Jacobs and Vêtements played with ChatGPT in the press notes for their latest shows.
During the latest Copenhagen Fashion Week, Ganni unveiled its own exploration of AI's impact on the fashion industry. Before starting to work on the brand’s Spring/Summer 2024 collection, Ganni's creative director Ditte Reffstrup pondered on the role of AI in the fashion industry with her husband, the brand's co-founder, Nicolaj.
On the brand’s site she states: "I was thinking about AI's potential for good. As a mother of three kids, and as a woman running a business, I think a lot about how to raise my kids to be kind, to feel secure, to be responsible."
Playing around with ChatGPT, Nicolaj asked the system to come up with a brief for Ganni's next collection. In the end, the design team didn't take into consideration the document generated by ChatGPT, but this interaction between a human and Artificial Intelligence ended up all the same informing the new collection, entitled "Hello, World!" the traditional phrase used to illustrate the very basic functionality of a programming language or demonstrate how to write a simple program.
Worried by AI, but also fascinated by technology, Ditte started wondering indeed if it would be possible to come up with a more ethical Artificial Intelligence, "kinder, more thoughtful (…) one that does good" and turned for help to Danish digital artist Cecilie Waagner Falkenstrøm.
With her studio ARTificial Mind, Waagner Falkenstrøm created a purpose-built Generative Pre-training Transformer (GPT) similar to Open AI's ChatGPT.
This AI was trained with a substantial amount of data, mainly brand data, online interviews, and social media interactions within the "Ganni Girls" community of fans and influencers.
During the show, guests could also engage with this AI, posing questions via microphones attached to living trees that talked via speakers hanging from their branches.
Placed around the venue, the trees represented humanized AI bots, gentle yet eerie, reassuring the audience at the beginning of the show with an encouraging message, "We come in peace".
"We wanted the spirit of the AI to be embodied by a physical form, something natural - my first thought was trees - they are such a symbol of hope and life," Ditte Reffstrup explained on the brand’s site.
The influence of Artificial Intelligence extended to creating the show's soundtrack that comprised favorites from the Ganni community in a playlist that included Daft Punk, Shakira and Blackpink.
Beyond the AI activations, Ganni introduced novel collaborations during the show, including its second partnership with New Balance, a debut collaboration with British eyewear label Ace & Tate and a five-piece collaboration with model Paloma Elsesser who opened the show.
While there was an interesting interplay between romantic dresses with floral motifs and sculpted suits, the collection (that also had some Miu Miu moments) was more intriguing for materials than for the actual garments (that included embellished jeans and practical designs with animal prints dedicated to urban tigresses; functional frocks and pants in metallic shades). For the next season Ganni employed indeed innovative materials such as Oleatex - a plant-based leather alternative created from olive oil production waste; Algreen's biodegradable, plastic-free sequins made from seaweed and agricultural waste, and Circulose, a recycled material made from 100% discarded textile waste.
But let's go back to the Artificial Intelligence system employed in this collection and ponder a bit on it as there are interesting points to make here. If the robots/living trees represented "Ganni girls", the brand's fans preoccupied by sustainability issues yet keen on living a carefree lifestyle, aren't we reducing the brand's consumer to two main things - their commitment to sustainability and their penchant for vibrant partying? Are we implying that this is all that can be said about the brand's consumers?
Besides, can we consider ethical to solely train an AI using favorable remarks about ourselves from our own community of fans? (actually, there may also be a copyright issue here as the AI tool was trained with interviews and features about Ganni - but were the people who wrote them asked if it was fine to use their material to train an AI?) Isn't this similar to those companies that mainly train their AI system with pictures of white and thin models?
Yes, it is logical that in your own fashion shows you just want an AI saying positive things about you. Yet this is actually where we enter a rather peculiar terrain: currently, investing in training an AI tool to speak positively about you might be comparable in cost to inviting to your event an influencer that can subsequently flood social media with images and praise in exchange for a sponsored trip and compensation. The costs are the same at the moment (if not higher, actually) as right now you still need to scrape data about your brand, create an AI tool, train it, etc.
However, the day will come when it will become cheaper to have an AI praising you than an influencer. Actually, more prominent, wealthy and powerful brands could employ multiple AI entities that could shower them with compliments, perhaps even generating clever comments and comparisons with past collections if the AI has been trained comprehensively.
As we move closer to resembling robots ourselves as fashion trends continuously steer us towards uniformity (think about this summer's pink trend caused by the "Barbie" movie), the prospect of training AI solely for positive endorsements raises concerns.
This could engender naïve and controlled systems, masquerading as ethical, while, in reality, we'd merely be feeding them the affirmations we want to hear. Such an AI might mimic the behavior of a paid influencer, adopting an ingratiating and docile stance - it would be a risk-averse AI, devoid of any critical perspective and afraid of expressing anything challenging or of uttering any negative remarks. In a nutshell, rather than training an AI for higher purposes, we would be crafting entities that cannot voice constructive criticism but can replicate a model consumer and fan, offering monotonously positive feedback.
Comments