Towards Sustainable Artificial Intelligence: A Comprehensive Review and Comparative Analysis of Deep Learning Models' Carbon Footprint

Research from arXiv highlights the growing concern regarding the environmental impact of Artificial Intelligence (AI) and Machine Learning (ML), particularly Deep Learning (DL) models. Despite their utility in automating complex tasks, the substantial computational resources required for these models lead to significant energy consumption and carbon emissions. This paper systematically reviews existing research on 'Green AI' and 'Green DL', focusing on optimization techniques designed to mitigate AI's environmental footprint.

Why it matters

The environmental sustainability of AI development and deployment is becoming a critical strategic consideration across all sectors. Organizations investing in AI must acknowledge and address the carbon footprint associated with these technologies to align with broader sustainability goals and regulatory expectations.

Key insights

• AI and ML technologies, especially Deep Learning architectures, have high energy demands.
• The energy demands of AI/ML contribute to significant carbon emissions.
• Large-scale model deployment exacerbates the environmental impact of AI.
• Research areas like 'Green AI' and 'Green DL' are emerging to address these concerns.
• Optimization techniques are being explored to reduce the environmental impact of AI models.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.09998