Unlocking marginal SDG dynamics: A KRLS machine learning analysis of AI technologies and solar energy

Abstract

Studies quantifying AI’s impact on Sustainable Development Goals (SDGs) often rely on proxies that inaccurately reflect AI progress. Moreover, focusing solely on environmental and growth indicators provides an incomplete picture of AI’s overall contribution to the SDGs, as the SDG framework encompasses a broader set of interconnected goals. Therefore, this study unveils the marginal impacts of AI and solar energy (SEN) directly on the SDG Index (SDGI) by using the Kernel-Based Regularized Least Squares (KRLS) machine learning approach for the 10 largest economies from 2000‒2022. While the study found an overall positive average marginal impact of AI on the SDG Index, indicating significant progress driven by AI technologies, the analysis across different quantiles revealed variability. Specifically, at the 25th quantile, AI appears to hinder SDG progress. This could be due to negative externalities from AI adoption, like its use in accelerating non-renewable energy production and resource-intensive consumption, or from countries’ insufficient technological application capabilities. However, at higher quantiles (likely representing countries with better SDG achievement and greater AI maturity), the marginal effects of AI become increasingly positive, suggesting its beneficial use in areas that support SDGs. Marginal effects of SEN on SDGI are found to be positive, showing a positive connection between SDGs’ achievements and solar energy adoption. The marginal effects of economic globalization (EGB) and institutional productive capacity (INP) on SDGI are found to be positive. Finally, policies to boost AI and solar energy adoption, as well as exploring potential applications of AI across various sectors for sustainable development, are discussed.