Integration of Artificial Intelligence in Taxation: Are Kenya and Tanzania Ready?

Abstract

Artificial intelligence (AI) is increasingly transforming tax administration through predictive compliance analytics, automated taxpayer services, fraud detection, audit selection, risk profiling, and the integration of previously fragmented datasets. While these innovations offer significant benefits for governments seeking to boost domestic resource mobilization, they also raise critical questions about institutional readiness, data quality, digital infrastructure, taxpayer rights, cybersecurity, algorithmic bias, administrative capacity, and public trust. This paper investigates whether Kenya and Tanzania, two of East Africa's largest and most reform-driven tax authorities, are prepared to adopt AI in tax administration.

Using a comparative qualitative approach, the analysis examines tax reform documents, digital government initiatives, international reports, and emerging AI governance frameworks. The findings indicate that both countries are at an advanced-emerging stage of AI readiness but occupy different strategic positions: Kenya has made the most progress, implementing an AI-based automated return validation system since January 2026, supported by a national AI strategy and proposed AI legislation. Tanzania has established key infrastructure, the Integrated Domestic Revenue Administration System (IDRAS), launched in February 2026 to unify tax processes, but lacks a clear AI enforcement plan.

The paper introduces the AI Tax Administration Readiness Framework (AITARF), which has eight dimensions. It emphasizes that digitalization alone isn't sufficient for AI readiness and recommends a phased approach: Digitize, Integrate, Govern, Analyze, Automate, and Apply AI, prioritizing AI as a complement to human officials rather than a replacement.

Keywords: artificial intelligence; taxation; tax administration; Kenya; Tanzania; domestic resource mobilization; digitalization; tax compliance; machine learning; algorithmic governance.

DOI: 10.7176/EJBM/18-8-08

Publication date: August 30^th 2026