Does artificial intelligence improve social equity in public services? A systematic literature review on opportunities and pitfalls

Abstract

Purpose – This paper explores challenges and solutions between artificial intelligence (AI) and social equity within the public sector. A cross-disciplinary synthesis explains where equity risks originate and which governance levers are most effective in mitigating them. Design/methodology/approach – The literature was examined across three databases (Web of Science, EBSCO and Scopus). A final corpus of 128 peer-reviewed articles (2016–January 2025) was qualitatively coded and analyzed for study type, AI modality, equity challenges and remedies aligned to governments' roles. Findings – Interest has expanded rapidly since 2021, with conceptual work still dominant but empirical studies growing. Scholarship disproportionately centers on regulatory and user roles, while enabler and leader roles remain underdeveloped yet pivotal for upstream capability distribution and downstream accountability. Research limitations/implications – The framework simplifies a complex governance reality and requires empirical validation; longitudinal and practitioner-engaged studies are needed to track long-term equity effects. Practical implications – The framework helps policymakers match equity risks to concrete instruments–regulatory audits and due process (regulator), human in the loop design and appeals (user), market shaping procurement and capacity building (enabler) and sovereign infrastructure and solidarity mechanisms (leader). Social implications – As AI transforms society, awareness of the complexities to manage and the potential solutions is a critical frontier to investigate and disseminate to support the government in the new digital age. Given the centrality of social equity for governmental organizations, a critical analysis of potential challenges and solutions is essential, particularly considering the complexity of government roles concerning AI adoption. Originality/value – The study integrates social equity theory with a role-based framework to map practical levers for equitable AI in government and to expose role interdependencies.