ARTIFICIAL INTELLIGENCE AND ENVIRONMENTAL SUSTAINABILITY: A FIRST LOOK INTO CITIZENS' UNAWARENESS

Artificial intelligence (AI) has significant potential to enhance resource efficiency and reduce environmental costs. Nevertheless, AI systems-especially large language models (LLM)-are energy-intensive and resource-demanding. The ecological cost stems not only from the energy required for training and inference, but also from the infrastructure needed to sustain AI operations, including data centers and cloud services. From a sociopsychological point of view, it is extremely relevant to understand what beliefs individuals hold on the relation between AI and sustainability. Surprisingly, no research has been conducted on the topic so far. Hence, we conducted a first exploration through 44 interviews in Italy. Results show that individuals held no significant beliefs about the relation between AI and sustainability, in terms of both perceived advantages and disadvantages. Hence, results suggest a widespread lack of awareness about AI's sustainability, thus highlighting a significant cognitive and communicative gap. Additional cross-cultural research is needed to address a significant research gap and inform policies designed to enhance citizens' awareness of the environmental costs and benefits associated with advanced AI applications. This appears essential for cultivating a more reflective and knowledgeable approach to Green AI.