How is AI as a fashion designer?
A traditional fashion designer works with techniques like sketching and draping. The emergence of AI platforms shifts the creative process to writing prompts, where linguistic precision is crucial. Marie Ledendal, Magdalena Petersson McIntyre, and Despina Christoforidou, researchers at Lund University, are now investigating how generative AI design platforms are transforming the field by mediating creative processes, reshaping design roles, and influencing aesthetic outcomes. The research team worked by writing prompts and then examining the results from four platforms based on two themes: what they call "the limits of democratization" and "the myth of the muse."
One question is who gets to design on these platforms.
"Users are invited to create, the tools are shaped by professional design assumptions. As a result, a dual positioning emerges, between accessibility for hobbyists and application for professionals, where users are equipped with generative tools, but left without guidance during critical phases of the design process." they write in the article "The Myth of the Muse." "Users are encouraged to create, but only up to the boundaries of the platform’s affordances, and any cultural and technical capital to act beyond those boundaries are expected to be possessed by the users themselves.."
Beyond access, democratization also concerns who is visible in the creative results, and this involves representation, visibility, and inclusion in terms of bodily norms. To explore these biases, prompts were used that specified age, body type, ethnicity, nationality, or varying physical ability.
"While the platforms present themselves as tools for personalisation and creativity, their defaults consistently favoured youthful, slender, light-skinned, and binary-gendered figures. (...) This underscores a persistent aesthetic bias and shows that algorithmic outputs often default to narrow ideals, even when instructed otherwise.."
Finally, the claim that AI functions as a creative partner or "digital muse" is also examined.
"Across the studied platforms, the outputs showed little of the surprise, or breadth of outcomes typically associated with the muse. Even after multiple prompt adjustments, as we discussed in the previous section, the systems 17 repeatedly generated young, slim, white, long-haired women in commercial styles. Rather than enabling unexpected outcomes, the platforms defaulted to narrow aesthetic templates, suggesting that the promise of endless creativity is constrained by algorithmic repetition and normative defaults."
The authors also argue that the platforms often lead to the disregard of important textile knowledge. "Prompting becomes the primary skill, while material properties, stretch, weight, and handle, are peripheral, or absent.."
Want to read the full article "The Myth of the Muse"? Here's the link: The myth of the muse: How AI fashion design platforms and prompting reshape creative agency | Intellect
The work is part of the Fashioning AI project, involving Magdalena Petersson McIntyre (project manager), Gabriella Nilsson from the Department of Arts and Cultural Sciences, and Despina Christoforidou from Industrial Design.