DECODING THE ENVIRONMENTAL FOOTPRINT OF SOFTWARE USING EXPLAINABLE AI

The rapid expansion of software systems and digital services has led to a noticeable rise in global energy 
usage, which in turn contributes to increasing carbon emissions and environmental concerns. Although 
modern computing technologies have greatly improved efficiency, speed, and scalability, the environmental 
impact of software applications often remains overlooked and insufficiently examined. As a result, the true 
ecological cost of software operations is not clearly visible to developers or decision-makers. This study 
introduces an Explainable Artificial Intelligence (XAI)–based framework designed to assess, evaluate, and 
interpret the environmental impact of software applications, focusing specifically on energy consumption and 
carbon footprint. 
The proposed framework combines energy monitoring tools, system performance indicators, and machine 
learning models to estimate environmental costs under varying software workloads. By incorporating 
explainable AI methods, the framework offers clear and understandable insights into how individual code 
segments, algorithms, and runtime behaviors influence overall energy consumption. This level of transparency 
helps developers identify energy-intensive operations and supports informed decision-making for developing 
more sustainable and energy-efficient software solutions. 
To validate the framework, a prototype system is implemented and tested using standard benchmark 
applications. The experimental results demonstrate the framework’s capability to deliver reliable energy 
predictions along with meaningful interpretability. The findings emphasize the role of explainable models in 
aligning software performance optimization with environmental sustainability. Overall, this research provides 
a practical pathway for advancing green software engineering practices and encouraging the development of 
environmentally responsible computing systems.