Advancing Sustainable AI: Balancing Performance and Carbon Emissions with System of Systems Theory

Kailash Alle

Abstract: Artificial intelligence (AI) can reduce and increase a company's carbon footprint. This study explores this complex relationship. AI techniques can be used to fight climate change, but training these models can consume a lot of energy. Here, we examine AI and carbon emissions as interconnected systems. We analyzed six machine learning models and calculated their carbon footprint. Our goal is to promote "sustainable AI" where environmental responsibility is considered throughout the development process, from data collection to using the model. By advocating for sustainable practices, we encourage the creation of effective AI that's also eco - friendly.

Keywords: Sustainable AI, Machine Learning, Carbon Emissions, Energy Efficiency, Green AI, Deep Learning