Green Recommender Systems – A Call for Attention

Recommender systems are pivotal in numerous industries, from e-commerce to streaming platforms, significantly influencing user choices and engagement. However, the computational costs of these systems have escalated sharply, contributing to growing environmental concerns. This paper advocates for the development of Green Recommender Systems—recommender systems designed to minimize environmental impact throughout their life cycle, from research and design to deployment and operation. While green systems strive to maintain the performance of traditional systems, they may accept trade-offs in metrics like accuracy to prioritize sustainability, often focusing on reducing energy consumption and CO2 emissions. The increasing complexity of machine learning models directly contradicts sustainability goals, and as such, a shift towards resource-efficient models is critical. We explore practical strategies to reduce the environmental footprint of recommender systems, such as using energy-efficient hardware and downsampling datasets. By incorporating sustainability into the design and evaluation of these systems, we aim to foster the development of algorithms that are both effective and environmentally responsible. This paper calls for the recommender systems community to expand the concept of performance to include energy efficiency, balancing accuracy with sustainability for the long-term viability of recommendation technologies.