Algorithm or creativity? Fashion’s defining dilemma in the age of AI

There was a time, not too long ago, when designers had to answer unavoidable questions every time they faced the gatekeepers of the press: What inspires you? Who do you design for? What is your ideal woman? Questions meant to check boxes and keep the conversation going, which are still used today.

Although now they seem like trick questions: in an era dominated by digital media, creative work is no longer necessarily the result of one’s own imaginative expression, and authorship has come under greater scrutiny. The algorithm is looming large.

It’s the invisible hand that guides current behaviors, customs, and relationships. It not only controls and decides the content consumed through platforms and social networks, but also acts as a digital muse. From computer codes programmed to predict and prioritize anything that might hold users’ interest to algorithms that analyze vast amounts of data — the most visited websites, the most clicked links, the most shared elements — to ensure the greatest enjoyment of the online experience.

They act as curators of our tastes and preferences, filtering noise and organizing what would otherwise be a chaos of information overload. But what initially appeared like a logical support system has begun to pose questions that are difficult to resolve. Are we free to express ourselves, or are we shaping our creativity to fit what we think will favor the algorithm? Does this almost unconditional surrender to its power foster innovation, or does it reinforce creative homogeneity?

The issue has grown more complex since generative artificial intelligence entered society. These programs have shaken things up by creating original content — whether text, images, video, audio, or even software — in response to a user’s prompt. It opens up a seemingly limitless artistic horizon. And from a productivity and economic standpoint, it is highly profitable.

It’s still early days, but there have already been many milestones. Fascinated by the potential of the metaverse — an immersive gaming-like experience accessible from any device—and NFTs (non-fungible tokens: unique digital files in formats like JPG, GIF, or even audio, designed for online trading), Tommy Hilfiger was one of the pioneers in incorporating generative AI into the creative process. In 2018, he launched the pilot project Reimagine Retail in collaboration with IBM and the Infor Design & Tech Lab at New York’s Fashion Institute of Technology (FIT). The algorithm was trained to scan the brand’s archives — both runway collections and product imagery — and from there, it generated patterns, colors, and styles in the Hilfiger aesthetic. These were then used to create a capsule collection with FIT students.

“There’s speculation about the impact of AI, but we’re committed to advancing its role not just to help the industry in terms of productivity, but also creativity,” says Bhaskar Pant, executive director of the MIT Professional Education program — an institution designer Norma Kamali attended a couple of years ago.

Kamali took one of the university’s online courses, and from there, began thinking about how to expand her legacy after 57 years in the industry. Maison Meta, a fashion design startup, made it easy: together, they developed an AI program trained to generate categories (dresses, swimsuits, T-shirts...) from written prompts — not images or sketches. That meant Maison Meta had to equip it with an extensive fashion vocabulary. The result, Hallucinations, was unveiled at the last New York Fashion Week.

“AI doesn’t have a pulse, it won’t replace human passion. But it can enhance creativity in ways we’re still beginning to understand,” says Kamali.

Maison Meta is also the driving force behind AI Fashion Week, launched in 2023 — a competition for young designers that has already spotlighted talents like Rachel Koukal, Dmitrii Rykunov, Matilde Mariano, and Spaniards Magno Montero and Esther Souto Prego (whose collection was inspired by Sargadelos ceramics). These emerging designers are using free tools such as Midjourney, Pika Labs, CapCut, and even ChatGPT or DALL·E 3. There are countless programs like these — some open-source — that generate styles, suggest color palettes, and create patterns (computerized pattern-making is not new, but AI greatly speeds up the process and calculates fabric use for optimal efficiency).

The system that has truly changed the game is AiDA, created in 2013 by fashion analytics expert Calvin Wong, a professor at Hong Kong Polytechnic University. Its generative algorithm takes rough sketches and raw ideas and transforms them into detailed designs.

Naturally, the major luxury conglomerates have developed their own systems, primarily used as production or purchasing assistants. Kering has Madeline, which optimizes the consumer shopping experience. LVMH, through its tech investment arm Aglaé Ventures, has invested around €300 million ($344 million) over the past year in five startups developing generative AI systems (Borderless AI, Proxima, Lamini, Photoroom, H Company). These are tools aimed at improving productivity, digital marketing, and graphic editing.

Whether any of its brands are using such systems as design tools remains unclear — especially after the group’s official statement during last year’s LVMH Data AI Summit in Paris: “While these technologies offer remarkable creative potential, making it possible to generate bespoke designs and personalized experiences, they also raise questions about authenticity, as well as responsibility.”

A survey by consultancy Deloitte found that 20% of fashion brands plan to incorporate generative AI into their design processes, a figure McKinsey raises to 25% in The State of Fashion 2025 report. Rumors that some brands are exploring the idea of using generative AI as a creative director only add fuel to the fire. H&M and Levi’s have started using AI for visuals and campaigns featuring virtual models, and G-Star has launched its first generative AI-created pieces in the capsule collection Denim Design Reimagined.

“The craftsmanship involved can only be done by human hands and with a passion that no machine can have,” said Guram Gvasalia in a press release written by ChatGPT to present Vetements’ Spring-Summer 2024 collection, which was described as an anti-AI.

With projected extra earnings for the fashion sector estimated between €150 and €270 million ($170-$310 million) over the next five years (according to McKinsey), the industry is clearly entering a high-stakes algorithmic game of thrones.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition