The AI dilemma: Is fashion heading towards an algorithmically driven repetitive loop or its next chapter?
Is artificial intelligence a new tool for creativity in fashion?
This article was originally written by Demi Karanikolaou in Greek for Harper’s Bazaar Greece. You can find it here.
Cyclical nature of fashion
2023 has undoubtedly been the year that AI has been discussed the most in a fashion context. Between the positive and negative comments, many expressed their fears that AI will halt creativity in this industry, due to its algorithmic ways and dataset source, which contains only pre-existing knowledge of things that have already been done. However fashion’s cyclical and repetitive tendencies are hardly a new phenomenon. Throughout history, dress makers and consumers have been drawing inspiration from previous decades and their various styles or even the body type en vogue over a certain period. It is then without a question, why we are seeing fashion trends that were initially popular over a certain time, return, often with a modern day twist that deems them relevant to a new era. But that does not mean that we never see innovation in the space. Indeed, the latters creeps into our life in one way or another, being more evident in certain decades over others. Jeans, mini skirts, pants for women, chain bags or even full aesthetics, such minimalism, were all new at some point. Specific decades, such as the 60s with their space age fashion and modern furniture, became legendary for their innovative vision. Generally speaking however, there is always a tendency to revisit the past every so often. Already in the 70s, fashion trends setters longed for a return to romanticism and 30s silhouettes. However, in the age of Tik Tok, this has been overexaggerated, as digital platforms & social media accelerated this process. The creation and death of trends happens at the speed of light, giving very little time for truly new shapes, styles and materials. It is then without surprise that Vanessa Friedman of the New York Times noted that the 2023 fashion scene was a pure mashup of revivals: the revival of Y2k trends, the skinny body types and that of white male creative directors coming once again to the forefront of the industry.
The era of Generative AI: a new tool for creativity?
In this already complex structure of fashion trend cycles, we have to also look at the role of Generative Artificial Intelligence (AI). Generally speaking, Generative AI is defined as a type of artificial intelligence system capable of generating text, images, or other media in response to prompts. These models learn the patterns and structure of their input training data, in order to generate new data that has similar characteristics. In its essence, AI is drawing conclusions or creating outcomes from a specific pool of data it has been fed - which in theory can even consist of the whole internet. Naturally, this has raised some questions regarding the originality of AI generated design. In truth, the AI tools used to create them, are already drawing their knowledge from a dataset of references that already exist - to the extent of some claiming that, even if there would be no exact duplicate to an AI design output, it would still not be 100% original.
In addition to that, since AI is purely driven by datasets and algorithms, it is likely that companies will inevitably feed it with the financial data and goals they would want to achieve from a business perspective - potentially at the expense of real creativity and wild imagination. This is by no means a new tendency in fashion and art, most designers and creative directors in business today design garments that are both referencing the work of other artists before them, as well as using merchandising data to increase the sellability of the end products and achieve commercial success. This is precisely why there are fears that AI’s algorithmical ways might accelerate these tendencies as its mass workplace and daily life adoption continues. Social media algorithms, email marketing communication and product development are already using some form of Artificial intelligence in the background. The public has started using AI via Chat GPT and MidJourney and AI studies have already been integrated in the curriculum of notable fashion schools, meaning it is only a matter of time for AI to become a priority for all fashion brands. The question remains: “Is fashion heading towards an algorithmically driven repetitive loop of sellable items and safe designs that will bring in easy profits and will AI eventually destroy creativity?” Some of the industry’s pioneers heavily disagree.
Looking at the bright side of AI
Even though AI might have several limitations compared to the human imagination in the creative fields, its algorithms can create and calculate in ways that the human mind cannot, drawing its conclusions from large sets of unstructured and unrelated data sources.
“Generally speaking, throughout history our industry has benefited from every time that people invented new tools. Simply put, it is good to disrupt the status quo and AI can help designers create more interesting fashion as a co-pilot to real human creativity” said CuteCircuit co-founders, Francesca Rosella and Ryan Genz. Cutecircuit, the brand that dressed Katy Perry for the 2010 MET gala, is no stranger to innovation. Founded in 2004 it is the world’s first wearable technology fashion maison. Now, the brand is famously propelling AI at the forefront in order to aid creativity and innovation.
“AI is allowing both consumers, as well as designers to be more experimental. The first, by being exposed to the more diverse and innovative - compared to those on instagram - AI generated pictures, resulting in them feeling more adventurous with their style and eventually also buying into more eclectic pieces. When it comes to designers and the designing process, AI is an incredible tool that allows various Ideas to materialise in front of the designers eyes faster than ever before. This means that a creator does not have to be stuck on their only idea for the day, when the second one might actually be better. Instead they can give AI the right instructions and produce hundreds of different styles, colorways or styling options, to test their ideas before then go all in to curate, edit and choose what they want to push further in the production process. This previsualisation stage, enhances creativity in the industry in our opinion, by allowing for multiple tweeks, improvements and iterations of a product, saving time, money and creative energy in the development stages. Yes, fashion can be repetitive and AI is fed a specific dataset. However, The fact that the ingredients are not new, does not mean that the outcome is also not. For example, poets use words that already exist, but actually come up with a poem that has it’s own unique meaning. The thing that differentiates real creativity from random tests is the intention of the creator. One can copy a style, or recreate products without understanding their meaning. The exact same is true for the outcomes of powerful AI servers such as MidJourney. They should be used with intention if trying to add something to the industry. ” added the CuteCircuit co-founders.
Ironically enough, despite much of the early conversation about ChatGPT revolving around the demise of writing and creativity, it appears that the very success of generative-AI searches is determined by the quality of precisely written “prompts”, or the written instructions given to AI in order to request an output.
“ Detailed prompts help teach AI to think and refine better, essentially allowing the designer to have a creative dialogue with the tool and test out which idea works best, while visualising his thoughts and avoiding repeated mistakes and repetitions. It becomes thus, a matter of fostering creativity. Always pushing yourself forward, opening your mind and trying harder and harder until something new and innovative is produced. AI allows some randomness to get inserted in the process that propels the thought process further. There are infinite possibilities for creativity“ says Francesca Rosella
On the sustainability front, it is also undeniable that AI is an incredible tool in trend forecasting. Indeed, AI can analyse all of social media pictures, in order to help companies become more sustainable, by finding the exact customer profile for every product and producing it only in the amounts and styles necessary, eventually lowering returns and ags emissions, improving products and customer satisfaction.
Looking into the future
Every new technology has a period of adapting and since these tools are still at their infancy, we have yet to see the long-tail effect of generative AI models. This means there are some inherent risks involved in using them—some known and some unknown. For example, at this stage the outputs generated by AI can sometimes be biassed by the mainstream ideas that human society has already flooded the internet with. Or they might also be entirely false, which is why it is crucial to have solid human quality control performed every time. This forms what we call “hybrid intelligence”, aka the perfect collaboration of humans and machines.
AI cannot replace human creativity, but it can be a tool that frees it from daily tasks and other concerns. Automation has always been used/adopted by industries in order to make more money quickly. But it seems that despite the cyclical nature of fashion, AI is actually aiding creativity instead of simply repeating old trends based on data.
With designers and merchandisers using AI as a tool that helps them with mechanical and monotonous tasks, one has to wonder if actually AI is inevitably going to place a higher value on human creativity and overall on humanistic studies. Indeed, skills like art curation, styling, knowledge of literature and having an “eye” for beauty might become more valued than they are today. AI can free time up for humans to do their research: fashion designers can spend days at a library, immersing themselves in the arts. This luxury is something that is currently missing from the industry: the research, the knowledge of art movements and music. Hopefully kids in future fashion schools will be able to shift their focus to that.
As Francesca Rosella put it: “I see Ai as a genie in a bottle: if you want to create innovative fashion, you have to have very ambitious wishes or in this case, prompts. You can ask AI for anything you want, you can try again and again, you can curate. You can wish to see the future of the industry”
