Academic study outlines framework for integrating AI tools into fashion design education

The pedagogical integration of artificial intelligence across university-level fashion design programmes requires structured institutional frameworks to support creative exploration while mitigating technical design deficits, according to a study published in the peer-reviewed journal Fashion and Textiles and reported by Bioengineer.org.

Conducted by researchers Hyosun An and Minjung Park, the qualitative investigation gathered insights from educators across institutions in South Korea and the United States to examine how algorithmic tools intersect with student challenges in concept generation, technical garment construction, and final critique.

The findings emerge as apparel design schools and commercial design studios navigate the rapid rollout of generative platforms such as Midjourney, DALL-E, and specialised Creativity Support Tools (CSTs). While these digital tools significantly lower barriers for students lacking advanced freehand illustration skills—enabling rapid concept prototyping, market data filtering, and early visual iterations—instructors noted persistent structural bottlenecks in physical execution.

Specifically, eight out of ten surveyed educators reported that learners consistently struggle with pattern drafting, garment anatomy, and closure placements when translating digital renderings into tangible prototypes. The absence of directly exportable flat-pattern data—unlike established apparel CAD and virtual prototyping suites such as CLO3D—remains a clear technical boundary for generative image software in production workflows.

Four-phase pedagogical model

To address these limitations, the article structured a four-phase pedagogical model spanning problem definition, ideation, prototype development, and evaluation. Rather than treating automated tools merely as end-product image engines, the study advocates deploying them as cognitive aids. In the early research stages, large language models and text-to-image prompting assist with mood board synthesis and market benchmarking, provided students possess domain-specific vocabulary. During prototyping, 3D simulation tools offer non-destructive fit testing, which reduces physical sample waste before fabric is cut.

However, educators cautioned that uncritical overreliance on automated systems risks design fixation and can lead students to bypass foundational research methods. Seven out of ten instructors warned that studio assessment rubrics must shift away from evaluating visual volume toward grading prompt literacy, process refinement, and critical intervention.

Furthermore, the study cautioned academic administrations against implementing outright software prohibitions, noting that blanket bans often trigger a "Streisand effect" that drives students toward unmonitored tools without guidance on copyright or ethical usage.

As the fashion supply chain increasingly demands digital proficiency alongside physical construction expertise, aligning design education with national occupational standards will require formal faculty training, institutional software subsidies, and updated grading frameworks. The researchers emphasized that subsequent investigations will need to track larger quantitative cohorts across European and global design academies to evaluate direct student learning outcomes and establish cross-market curricular standards.