Generative Artificial Intelligence and digital sobriety: for an ethical and responsible approach

Abstract

Generative Artificial Intelligence (GIA) is driving innovation across sectors, from art and medicine to the automation of complex tasks, by generating sophisticated texts, images, and simulations. However, this technological progress brings significant environmental concerns, primarily due to the high energy consumption and substantial carbon footprint associated with model training and inference processes. As digital sobriety becomes a priority to mitigate environmental impact, GIA faces a paradox: its large-scale deployment often contradicts sustainable practices. This communication explores ethical considerations surrounding GIA, focusing on strategies for reducing its ecological footprint while supporting responsible digital use