The 21st International Conference “Man and Working Environment”
SAFETY ENGINEERING & MANAGEMENT SCIENCE, INDUSTRY, EDUCATION (SEMSIE 2025)   
PROCEEDINGS OF PAPERS 
25-26 September 2025, SOKOBANJA, SERBIA  

Mirjana Milutinović , Mladena Lukić 

ORIGINAL SCIENTIFIC PAPER

HOW GREEN IS YOUR ALGORITHM? ASSESSING THE CARBON FOOTPRINT OF MACHINE LEARNING

DOI: 10.46793/SEMSIE25.087M
DOWNLOAD FULL PDF
Abstract:

The rapid advancement of machine learning (ML) and its extensive application across various fields have led to numerous innovative uses. However, large-scale ML systems require significant computational resources, energy usage, and result in associated carbon emissions, which have raised some concerns. ML can combat climate change with smart decision-making, but energy-intensive models like deep learning also impact the environment. In this paper, we applied the CodeCarbon, an open-source tool for estimating energy consumed and carbon dioxide (CO₂) emissions during the runs of ML models (Linear Regression, k-Nearest Neighbors Regressor, and Decision Tree Regressor). Both default and optimized models show low CO₂ emissions, with optimization resulting in slightly higher values. The impact of geographical locations related to the carbon intensity of electricity generation on emissions is also examined, along with the effects of utilizing the complimentary cloud service GoogleColab. Due to low emissions, applied ML algorithms are suitable for education, research, and practice. The increasing use of artificial intelligence (AI) highlights tracking of carbon emissions, even in lightweight ML algorithms, to introduce sustainable AI practices. The aim of this paper is to raise awareness of the energy and environmental cost of AI at all levels of research.

Keywords:

Sustainable AI, carbon footprint, machine learning models, CodeCarbon, GoogleColab

ACKNOWLEDGEMENTS:

This paper is supported by the Ministry of Science, Technological Development and Innovation of the Republic of Serbia pursuant to agreement № 451-03-137/2025-03/200148, goal 13

REFERENCES:
  • ALLEA (2022) Towards Climate Sustainability of the Academic System in Europe and Beyond. Berlin. ISBN: 978-3-9823967-1-2. DOI 10.26356/climate-sust-acad. https://allea.org/wp-content/uploads/2022/05/ALLEA-Report-Towards-Climate-Sustainability-of-the-Academic-System.pdf (accessed: May 2025)  
  • Altman, N. S. (1992). An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression. The American Statistician, 46, pp. 175–185. https://doi.org/10.1080/00031305.1992.10475879 
  • Alpaydın, E. (2010). Introduction to machine learning (2nd ed.). MIT Press. ISBN 978-0-262-01243-0 
  • Alzoubi, Y. (2024). Green artificial intelligence initiatives: Potentials and challenges. Journal of Cleaner Production. 468, 143090. https://doi.org/10.1016/j.jclepro.2024.143090 
  • Bannour, N., et al. (2021). Evaluating the carbon footprint of NLP methods: a survey and analysis of existing tools. In Proceedings of the Second Workshop on Simple and Efficient Natural Language Processing, pages 11–21. https://aclanthology.org/2021.sustainlp-1.2/ 
  • Bergstra, J., et al. (2011). Algorithms for Hyper-Parameter Optimization. Advances in Neural Information Processing Systems. 24. Eds J. Shawe-Taylor, R. Zemel and P. Bartlett, F. Pereira, K. Q. Weinberger. https://proceedings.neurips.cc/paper_files/paper/2011/file/86e8f7ab32cfd12577bc2619bc635690-Paper.pdf  
  • Bouza, G., et al. (2023). How to estimate carbon footprint when training deep learning models? A guide and review. Environmental Research Communications. 5, 115014. XXXXhttps://iopscience.iop.org/article/10.1088/2515-7620/acf81b CodeCarbon.  
  • https://codecarbon.io/ (accessed: March 2025) 
  • https://mlco2.github.io/codecarbon/index.html (accessed: March 2025) 
  • Dayarathna, D., et al. (2015). Data center energy consumption modeling:A survey. IEEE Communications Surveys & Tutorials, 18, pp. 732–794. DOI: 10.1109/COMST.2015.2481183.  
  • Dodge, J., et al. (2022). Measuring the Carbon Intensity of AI in Cloud Instances. Proceedings FAccT '22: Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency June 21 - 24. https://doi.org/10.1145/3531146.3533234.  
  • Electricity Maps. https://www.electricitymaps.com/ 
  • Ember (2025); Energy Institute - Statistical Review of World Energy (2024). https://ember-energy.org/latest-insights/global-electricity-review-2025/(accessed: May 2025)  
  • Freedman, D. (2009). Statistical Models: Theory and Practice 2nd. Cambridge University Press. https://doi.org/10.1017/CBO9780511815867. 
  • Google Colab, https://colab.research.google.com/(accessed: March 2025) 
  • Guo, B., et al. (2022). Energy‑efficient database systems: A systematic survey. ACM Computing Surveys, 55, 1–53. https://dl.acm.org/doi/10.1145/3538225 
  • Henderson, P., et al. (2020). Towards the Systematic Reporting of the Energy and Carbon Footprints of Machine Learning. Journal of Machine Learning Research 21 pp. 1-43. https://jmlr.org/papers/v21/20-312.html.  
  • The International Energy Agency (IEA). https://www.iea.org/energy-system/buildings/data-centres-and-data-transmission-networks?mod=livecoverage_web&utm_source=chatgpt.com(accessed: May 2025) IPCC. (2023). Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (eds.)]. IPCC, Geneva, Switzerland, 184 pp., https://www.ipcc.ch/report/ar6/syr/downloads/report/IPCC_AR6_SYR_FullVolume.pdf  
  • James, G. et al. (2013). An introduction to statistical learning 112. New York: Springer. ISBN 978-1-0716-1417-4 https://doi.org/10.1007/978-1-0716-1418-1 
  • Kaack, L. H., et al. (2022). Aligning artificial intelligence with climate change mitigation. Nature Climate Change 12, pp. 518–527. https://doi.org/10.1038/s41558-022-01377-7  
  • Lacoste A, et al. (2019). Quantifying the carbon emissions of machine learning. https://arxiv.org/abs/1910.09700
  • Lannelongue, L., et al. (2021). Green Algorithms: Quantifying the Carbon Footprint of Computation. Advanced Science. 8, 2100707. https://doi.org/10.1002/advs.202100707  
  • Lannelongue, L., & Inouye, M. (2023). The carbon footprint of computational research. Nature Computational Science 3, 659. https://doi.org/10.1038/s43588-023-00506-2  
  • LEAFframeworkTake part in LEAF | Sustainable UCL - UCL –University College London https://www.ucl.ac.uk/sustainable/case-studies/2020/aug/take-part-leaf (accessed: May 2025)  
  • Luccioni, A., et al. (2023). Estimating the carbon footprint of BLOOM, a 176B parameter language model. The Journal of Machine Learning Research. 24 pp. 11990 - 12004 https://dl.acm.org/doi/10.5555/3648699.3648952.  
  • Olawade, D., et al. (2024)a. Artificial intelligence in environmental monitoring: Advancements, challenges, and future directions. Hygiene and Environmental Health Advances, 12, Article 100114. https://doi.org/10.1016/j.heha.2024.100114  
  • Olawade, D., et al. (2024)b. Artificial intelligence potential for net zero sustainability: Current evidence and prospects. Next Sustainability, 4, 100041. https://doi.org/10.1016/j.nxsust.2024.100041.  
  • Patterson, D. et al., (2021). Carbon Emissions and Large Neural Network Training. Machine Learning. https://doi.org/10.48550/arXiv.2104.10350. 
  • Rolnick, D., et al., (2019). Tackling climate change with machine learning. ACM Computing Surveys. 55 pp. 1-96. https://doi.org/10.1145/3485128 
  • Schwartz, R. et al., (2020). Green AI. Communications of the ACM, 63, pp. 54-63. https://doi.org/10.1145/3381831. 
  • Strubell, E., et al. (2019). Energy and Policy Considerations for Deep Learning in NLP. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, p. 3645–3650, Florence, Italy. https://aclanthology.org/P19-1355/.  
  • Strubell, E., et al. (2020). Energy and policy considerations for modern deep learning research. Proceedings of the AAAI Conference on Artificial Intelligence, 34, pp. 13693-13696. https://doi.org/10.1609/aaai.v34i09.7123.  
  • Thangam, D., et al. (2024). Impact of Data Centers on Power Consumption, Climate Change, and Sustainability. In K. Kumar, V. Varadarajan, N. Nasser, & R. Poluru (Eds.), Computational Intelligence for Green Cloud Computing and Digital Waste Management (pp. 60-83). IGI Global Scientific Publishing. https://doi.org/10.4018/979-8-3693-1552-1.ch004  
  • UC Irvine Machine Learning Repository site https://archive.ics.uci.edu/dataset/320/student+performance (accessed: March 2025) 
  • Vinuesa, R., et al. (2020). The role of artificial intelligence in achieving the Sustainable Development Goals. Nature Communications, 11, p. 233. https://doi.org/10.1038/s41467-019-14108-y  
  • Wang, J., et al. (2022). Sustainable AI: Environmental implications, challenges and opportunities. Machine Learning and Systems, 795–813. https://arxiv.org/abs/2111.00364.  
  • Xing, E. & Monck, A. (2023). AI emissions are fueling a new doomerism, this time it’s climate change. https://fortune.com/2023/12/12/ai-emissions-doomerism-climate-change-environment-tech/  
  • Wu, C.‑J., et al. (2022). Sustainable AI: Environmental implications, challenges and opportunities. Machine Learning and Systems. https://arxiv.org/abs/2111.00364  
  • SEDE Impact – Usage Statistics. (n.d.). Retrieved May 2025, from https://portal.xsede.org/#/gallery (accessed: May 2025)

© 2023 Faculty of Occupational Safety - Multimedia