Causal Factors and Core Categories of Artificial Intelligence Ethics in Accounting

Abstract

This study aimed to identify and explain the causal factors and core category of artificial intelligence ethics in accounting and to empirically examine the effect of causal factors on the ethical challenges of AI-based accounting. An exploratory sequential mixed-methods design was employed. In the first stage, 124 studies related to the applications and capabilities of artificial intelligence in accounting were reviewed and the initial components were extracted. The identified capabilities were then prioritized using the fuzzy analytic hierarchy process and expert judgments. In the qualitative phase, semi-structured interviews were conducted with 22 experts in accounting, auditing, and financial information technology selected through purposive sampling. Qualitative analysis was restricted to identifying causal factors and the core category. In the quantitative phase, a questionnaire developed from the qualitative findings was administered to accounting professionals, auditors, academics, and relevant technology specialists, and 212 valid responses were analyzed using partial least squares structural equation modeling. Structural model results indicated that causal factors had a positive and statistically significant effect on the core category of ethical challenges in AI-based accounting (β=0.542, t=7.97, p<0.001). The coefficient of determination for the core category was 0.294, indicating that the causal factors explained 29.4% of its variance. The effect size was 0.417, representing a strong effect. The predictive relevance index was Q²=0.182, demonstrating acceptable predictive capability. In addition, the SRMR value of 0.047 indicated satisfactory overall model fit. The findings demonstrate that ethical challenges associated with AI in accounting arise from identifiable causal factors, including lack of algorithmic transparency, limited explainability, inherent system complexity, pressure for rapid technology adoption, and deficient or biased data. These factors significantly intensify ethical concerns involving transparency, accountability, privacy, algorithmic bias, and professional independence.