Sustainable AI: Reducing Environmental Impact

The field of artificial intelligence is shifting towards a more sustainable and environmentally conscious direction. Researchers are focusing on developing energy-efficient models, reducing carbon footprint, and promoting green AI principles. This trend is driven by the growing concern about the environmental impact of large language models and other AI systems. Studies have shown that small language models can be a more sustainable alternative, and that careful design of prompts can guide these models towards greener software development. Additionally, there is a need for energy-aware scheduling, cost optimization, and physical host-independent energy estimates in virtualized environments. Noteworthy papers include: Generating Energy-Efficient Code via Large-Language Models, which empirically assesses the energy efficiency of Python code generated by LLMs against human-written code. Green Recommender Systems: Understanding and Minimizing the Carbon Footprint of AI-Powered Personalization, which provides guidelines for researchers and practitioners on how to minimize the environmental footprint of their work and implement green recommender systems.

Sources

Benchmarking Energy Efficiency of Large Language Models Using vLLM

Toward Green Code: Prompting Small Language Models for Energy-Efficient Code Generation

Data-Driven Energy Estimation for Virtual Servers Using Combined System Metrics and Machine Learning

Generating Energy-Efficient Code via Large-Language Models -- Where are we now?

Green Recommender Systems: Understanding and Minimizing the Carbon Footprint of AI-Powered Personalization

Built with on top of