An Assessment of Economic Public Policy in Algeria During the Period 2015–2024

Abstract

This study examines the complex relationship between artificial intelligence (AI) and environmental sustainability, focusing on the paradox wherein technology—despite reducing resource consumption—relies on infrastructure that consumes increasing amounts of electricity, water, and other resources. The study argues that the environmental value of AI should be measured not merely by predictive accuracy but by its net environmental benefit, accounting for computational costs, impacts on climate and water, data-related risks, and risks associated with algorithmic decision-making processes. Employing a descriptive, analytical, and comparative approach, the study analyzes academic literature, international reports, and regulatory frameworks, with a particular focus on the European Union, UNESCO, the OECD, and Saudi Arabia.

The study reveals that machine learning and deep learning offer significant potential for monitoring air and water quality, analyzing satellite imagery, predicting floods and droughts, enhancing energy efficiency, and preserving biodiversity. However, realizing this potential depends on data quality, the model's contextual suitability, and the ability to interpret and validate results. The research concludes that the principles of "Green AI" must evolve from a mere technical slogan into a regulatory standard encompassing the disclosure of energy and water consumption and emissions, the selection of the most efficient models, and continuous system re-evaluation throughout the lifecycle. Finally, the study proposes a three-stage governance model that integrates technical, institutional, and regulatory levels with cross-cutting principles—such as continuous auditing and human oversight—that apply to all processes.