LLMGreenRec: LLM-Based Multi-Agent Recommender System for Sustainable E-Commerce

A team of researchers from Vietnam has published a paper on LLMGreenRec, a novel multi-agent recommender system designed to bridge the gap in sustainable e-commerce. The system leverages Large Language Models (LLMs) to move beyond traditional session-based models, which are often optimized for short-term conversions and fail to capture nuanced user intent for eco-friendly products. LLMGreenRec employs a framework of specialized AI agents that collaboratively analyze user interactions and iteratively refine prompts to deduce green-oriented intents, thereby prioritizing recommendations for sustainable products.

This intent-driven, multi-agent approach offers a dual benefit: it actively guides users toward more environmentally conscious purchases while also aiming to reduce the system's own digital carbon footprint. By minimizing unnecessary user interactions and the associated computational load, the framework addresses the energy consumption often criticized in large-scale AI systems. The research, accepted for the Proceedings of the Conference on Digital Economy and Fintech Innovation (DEFI 2025), validates LLMGreenRec's effectiveness through extensive experiments on benchmark datasets, presenting it as a robust technical solution for fostering a more responsible digital economy.