<![CDATA[This research examines the application of artificial intelligence (AI) in modern tax administration, focusing on three domains: fraud detection, taxpayer services, and operational efficiency. The main problems identified are the gap between AI's rapid technical advancement and the readiness of governance, legal, and ethical frameworks, as well as the lack of integrated analysis across these dimensions. The objective is to synthesise empirical evidence on AI adoption opportunities and challenges, and to formulate actionable policy recommendations. Using a systematic literature review with thematic analysis of 20 sources (2021–2026), including OECD reports, peer-reviewed articles, and institutional studies, this research compares evidence from multiple jurisdictions (EU, US, China, Indonesia). The main findings show that AI adoption soared from 9% (2016) to 69% (2023), with fraud detection rates reaching 50%, potential efficiency gains of up to 40%, and processing time reductions of up to 70%. However, five critical challenges are identified: algorithmic bias and opacity, legal ambiguity, data privacy risks, accountability gaps, and human resource constraints. The synthesis reveals a non-linear (inverted U‑shape) relationship between AI usage intensity and professional vigilance. This research concludes that AI offers substantial opportunities, but success depends on humane, transparent, and accountable governance. An integrative three‑phase roadmap (inception, consolidation, optimisation) is proposed to balance power and trust dimensions in AI‑enabled tax administration.]]>
Copyrights © 2026
