The Role of Artificial Intelligence in Poverty Governance: A Systematic Literature Review of Innovations and Implementation Challenges

Artificial intelligence (AI) is increasingly embedded in development systems, enabling new capabilities for poverty prediction, social protection targeting and service delivery optimisation. However, the implications of these technologies for poverty governance—the institutional mechanisms for designing and delivering poverty reduction strategies—remain fragmented in the literature. This study conducted a PRISMA 2020-guided systematic review of peer-reviewed journal articles and scholarly book chapters published between 2015 and 2025 and retrieved from Scopus, Web of Science and DOAJ. Following title/abstract screening, full-text eligibility assessment and quality appraisal, 48 studies were selected, thematically identifying cross-cutting patterns related to system performance, implementation processes, governance considerations and contextual constraints. The reviewed literature indicates that AI can improve poverty governance through multimodal data integration, enhanced targeting accuracy and automated administrative processes. However, persistent challenges include biased datasets, infrastructural limitations, regulatory gaps and ethical risks such as algorithmic bias and digital exclusion, which may reinforce structural inequalities. The review contributes an integrated evidence base and introduces a conceptual framework for understanding AI in poverty governance, highlighting that developmental gains depend on robust data governance, inclusive digital infrastructure, context-sensitive design, algorithmic transparency and institutional capacity. Future research should prioritise impact evaluation, fairness-aware AI, participatory design and scalable approaches for low-resource environments.