Research on AI Computational Energy Consumption Optimization and Green AI
Jiawei Xu · SHS Web of Conferences · 2025
With the rapid development of Artificial Intelligence (AI), many deep learning models have been extensively applied in various fields, and correspondingly, the environmental issues arising from the development of AI have become one of the pressing arguments at present. Due to the significant impact on the environment caused by the high energy consumption and high emissions of AI, it is crucial to figure out how to reduce the energy costs and environmental costs of AI. This study mainly aims to promote the development of green AI by recognizing, analyzing, and improving AI consumption. Firstly, it examines the sources of energy consumption and carbon footprint of AI, which stem from model training, data storage, and data transmission. Subsequently, it optimizes the energy consumption of AI computing, covering both software and hardware aspects. Moreover, it also elaborates on the applications of green AI in specific fields. Finally, it discusses the existing challenges and future trends of green AI.