What is it about?
Software sustainability is about more than energy use: it also covers social and economic aspects. Machine learning (ML) systems are now everywhere, but we know little about how the people who build them deal with sustainability in practice, beyond reducing the carbon footprint. In this study, we asked ML practitioners directly. We first interviewed eight experienced ML engineers, and then ran a survey with 203 ML practitioners. We looked at how they describe sustainability, which software engineering practices they use to support it, and which challenges stop them from doing more.
Featured Image
Photo by Chris Ried on Unsplash
Why is it important?
We found a clear gap between awareness and practice: practitioners know that sustainability matters, but they rarely manage it in a systematic way. This points to concrete needs—clearer guidelines, ways to measure sustainability, and regulatory support—that researchers, companies, and policy makers can work on.
Perspectives
Sustainability is often reduced to the energy needed to train a model, but the engineers we involved described it in a broader way—environmental, social, and economic. What I take from this study is the gap between awareness and practice: practitioners know that sustainability matters, but they rarely manage it in a systematic way. I hope this work helps researchers and companies build the guidelines, metrics, and regulations that are still missing.
Prof. Stefano Lambiase
Aalborg Universitet
Read the Original
This page is a summary of: Sustainability of Machine Learning-Enabled Systems: The Machine Learning Practitioner’s Perspective, ACM Transactions on Software Engineering and Methodology, June 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3777553.
You can read the full text:
Contributors
The following have contributed to this page
