How artificial intelligence reshapes sustainable apparel design: structured evidence, multisource empirical anchoring, and an integration framework

Safety in Extreme Environments · Springer Nature · 2026-11-05 · DOI: 10.1007/s44498-026-00215-9

Аннотация

Artificial intelligence (AI) supports apparel concept generation, sewing-pattern engineering, virtual prototyping, production planning, consumer evaluation, and circular information systems. This study examines when these capabilities support sustainability-relevant decisions without compromising feasibility, aesthetics, market viability, or accountability. A structured narrative review organizes evidence by lifecycle stage, sustainability problem, and validation maturity. Four separately analyzed modules provide empirical anchoring: 4200 garment-design records; criterion-specific assessments by 28 specialists of 350 designs; a 3 × 2 × 2 consumer experiment with 864 recruited and 858 retained participants; and 426 enterprise pattern records with measured cutting-waste rates. The reported consumer cell means suggest higher purchase intention for co-created stimuli and labeled products, and lower intention at the higher price premium. Approximate Type-II ANOVA was reconstructed from the reported participant-level cell counts, means, and SDs; its precision is limited by rounding of those inputs. Within the observed enterprise, quarter, and product mix, ridge regression had reported test RMSE = 0.049 and MAE = 0.038, compared with baseline RMSE = 0.096. The AI-Sustainable Apparel Integration (ASAI) framework combines three structural pillars—data and technical integration, circular design performance, and validation and governance—with four evaluation dimensions: digital plausibility, design feasibility, market viability, and verified sustainability performance. Direct sustainability verification is limited to production-stage cutting waste; broader lifecycle outcomes require separate validation.

Читать у издателя →

Похожие статьи