Determinants and Challenges of Artificial Intelligence Adoption in Public Sector Governance: A Systematic Review
TL;DR
The TOE-I structure developed here gives policymakers and public sector managers a shared vocabulary for diagnosing where their own agencies stand and where intervention would matter most, and suggests that AI adoption in government is less a technology problem than an organisational and human one.
Abstract
Governments continue to invest heavily in Artificial Intelligence, yet uptake across public sector organisations remains uneven and often stalls well short of policy targets. This review examines why, drawing on 68 empirical studies published between January 2020 and December 2025 and screened under PRISMA 2020. An integrated Technology-Organisation-Environment-Individual (TOE-I) framework, combining TOE with the Unified Theory of Acceptance and Use of Technology (UTAUT), was used to organise the determinants and challenges reported across this body of work. Data quality and availability, leadership commitment, and trust in AI systems emerged as the most consistently cited determinants, alongside system compatibility, digital literacy, and interoperability. On the challenges side, organisational barriers dominate the corpus, ahead of technological, individual, and environmental barriers with legacy system integration, bureaucratic inertia, algorithmic bias, cultural resistance, and regulatory uncertainty recurring across contexts. Read together, these patterns suggest that AI adoption in government is less a technology problem than an organisational and human one: the tools exist, but the institutional conditions to use them well often do not. The TOE-I structure developed here gives policymakers and public sector managers a shared vocabulary for diagnosing where their own agencies stand and where intervention would matter most.