<![CDATA[The convergence of artificial intelligence and sustainable finance represents a fundamental transformation in investment decision-making, yet empirical evidence concerning effectiveness remains fragmented across diverse research domains. This study synthesises evidence from 43 peer-reviewed investigations spanning 2020 to 2024, examining artificial intelligence applications in environmental, social, and governance investing through systematic meta-analysis following PRISMA 2020 guidelines. Random-effects models demonstrate that artificial intelligence technologies significantly enhance risk-adjusted financial returns (standardised mean difference = 0.58; 95% confidence interval: 0.44-0.72; p<0.001), translating to approximately 5.2 per cent annual performance improvement, and environmental, social, and governance prediction accuracy (standardised mean difference = 0.53; 95% confidence interval: 0.38-0.68; p<0.001), representing 15 per cent error reduction compared with traditional methodologies. Ensemble machine learning demonstrates robust performance (standardised mean difference = 0.64; I²=45 per cent), whilst deep learning exhibits highest effects with substantial variability (standardised mean difference = 0.71; I²=68 per cent). Implementation success depends critically on data quality infrastructure (identified in 88 per cent of studies) and phased deployment strategies (effective in 64 per cent of cases). Moderate evidence certainty supports that artificial intelligence represents genuine capability advancement, though unexplained heterogeneity (I²=58-62 per cent) limits precise outcome prediction in specific contexts. Findings provide evidence-based guidance for investment managers adopting artificial intelligence technologies, policymakers developing regulatory frameworks, and researchers identifying future research priorities. Keywords: artificial intelligence; sustainable finance; ESG investing; meta-analysis; machine learning; investment decision-making; financial technology]]>
Copyrights © 2025
