AI RELIANCE AND PROFESSIONAL JUDGMENT QUALITY AMONG GOVERNMENT AUDITORS IN ACEH: THE MODERATING ROLE OF PROFESSIONAL SKEPTICISM

Abstract

The adoption of artificial intelligence (AI) in government audit processes, ranging from data analytics to algorithm-based anomaly detection, promises significant efficiency gains, yet raises global concerns regarding automation bias, the tendency of auditors to accept AI output without adequate critical evaluation. This study aims to analyze the effect of AI reliance on the professional judgment quality of government auditors in Aceh, and to test and explore the moderating role of professional skepticism in this relationship. This study employs an explanatory sequential mixed methods design, beginning with a quantitative phase through a survey of 110 government auditors from the Aceh Inspectorate, Regency/City Inspectorates, the Aceh Representative Office of BPKP, and the Aceh Representative Office of BPK RI, analyzed using Moderated Regression Analysis (MRA), followed by a qualitative phase through in-depth interviews with ten selected auditors. The (illustrative) quantitative results reveal a counter-intuitive finding: AI reliance has a significant negative direct effect on professional judgment quality, yet professional skepticism significantly moderates this relationship by buffering (weakening) the negative effect. The qualitative phase reveals the underlying mechanism: auditors with high trait skepticism treat AI as a complementary tool that is still cross-verified, whereas auditors with low skepticism tend to treat AI as a substitutive tool whose output is accepted uncritically, a pattern informants themselves described as 'thinking laziness resulting from excessive trust in machines.' This study offers novelty as one of the first studies to empirically examine the specific construct of AI reliance, rather than general audit technology use, within the context of public-sector auditors in a special-autonomy region such as Aceh, while extending automation bias theory by positioning professional skepticism as a cognitive buffering mechanism. The practical implications underscore the urgency of strengthening professional skepticism training integrated with AI literacy for government auditors, so that the digital transformation of auditing does not come at the expense of professional judgment quality, which remains the core of the audit profession itself.