Engineering Personalized AI Systems for Sustainable Fashion E-Commerce

Abstract

The rapid growth of fashion e-commerce has intensified sustainability challenges related to overconsumption, high return rates, and logistics-driven environmental impacts. In this context, personalized artificial intelligence (AI) systems offer a promising mechanism for aligning consumer preferences with sustainable consumption outcomes. This study investigates the engineering of personalized AI systems for sustainable fashion e-commerce by integrating consumer behavior signals, product sustainability attributes, and system-level control parameters within a unified personalization framework. Using a system-oriented analytical approach, the study evaluates personalization performance, sustainability outcomes, variable contributions, and user-segment responsiveness. The results demonstrate that sustainability-aware personalization improves recommendation diversity, fit confidence, and long-term engagement while significantly reducing return intensity and increasing exposure to low-impact fashion products. Interaction and temporal analyses further reveal that sustainability effects emerge cumulatively through repeated user–system interactions rather than instantaneously. The findings highlight the importance of embedding sustainability as an intrinsic design objective in AI-driven personalization systems and provide actionable insights for engineering responsible, scalable, and environmentally aligned fashion e-commerce platforms.