Towards Explainable AI for Government Workers

Explainable Artificial Intelligence (XAI) is essential for transparency, fairness, and accountability in government decision-making. Unlike commercial or healthcare AI, government AI operates in a high-stakes regulatory and political environment, directly impacting public trust and democratic governance. However, the lack of explainability in these systems poses serious risks, enabling manipulation, erosion of civil service integrity, and reduced accountability. Recent critiques of anti-expertise movements highlight how AI can either enhance democratic governance or obscure accountability. Given its role in national security, social services, and policymaking, XAI is not just a technical necessity but a fundamental safeguard against unchecked automation and political misuse. Explainable AI ensures that civil servants, policymakers, and citizens can trust and contest AI-driven decisions, reinforcing fairness, oversight, and agency. This paper argues for the urgent need to develop XAI frameworks tailored to government applications. We examine how government AI differs from other domains, why existing XAI research falls short, and how new frameworks can mitigate the risks of opaque AI in public institutions. By prioritizing explainability, we ensure AI serves the people, upholds democratic integrity, and strengthens—rather than undermines—public governance.