An Ethical Audit Model for the Governance of Indonesia’s Digital Tax System: Integrating the Principles of Fairness, Transparency, and Accountability

Abstract

The transformation of the digital tax system through the implementation of the Core Tax Administration System (CTAS) marks a new era in artificial intelligence (AI) based fiscal governance in Indonesia. While bringing high efficiency, AI implementation also raises ethical challenges such as algorithmic bias, lack of transparency in automated decisions and unclear human responsibility for system outcomes. This study aims to develop an AI Ethics Audit Model that integrates the principles of fairness, transparency and accountability (FTA) to strengthen fair and accountable digital tax governance. The research approach used is a qualitative descriptive approach with a development research model design, which combines a literature review of 45 journals (2018–2025), in-depth interviews with five experts (three DGT officials and two AI ethics academics), and analysis of policy documents such as PMK No. 81 of 2024 and ISO/IEC 42001:2023. The results show that there is still a risk of data bias, limited algorithm documentation and the absence of a formal ethics oversight mechanism at the DGT. Based on these findings, five stages of an AI ethics audit are formulated: audit planning, ethical risk identification, control evaluation, audit reporting and monitoring and follow up. This model provides a theoretical contribution by strengthening the concept of ethical assurance in digital governance, and a practical contribution by providing an ethical oversight framework for the Directorate General of Taxes (DGT) to ensure the transparent, fair and responsible implementation of AI in the national tax system.