Abstract
The latest advancements in machine learning, specifically in foundationmodels, are revolutionizing the frontiers of existing software engineering (SE)processes. This is a bi-directional phenomona, where 1) software systems arenow challenged to provide AI-enabled features to their users, and 2) AI is usedto automate tasks within the software development lifecycle. In an era wheresustainability is a pressing societal concern, our community needs to adopt along-term plan enabling a conscious transformation that aligns withenvironmental sustainability values. In this paper, we reflect on the impact ofadopting environmentally friendly practices to create AI-enabled softwaresystems and make considerations on the environmental impact of using foundationmodels for software development.