The Impact of AI-Powered Audit Tools on Audit Quality and Fraud Detection in Nigerian Deposit Money Banks

Abstract

This study examined the impact of AI-powered audit tools on audit quality and fraud detection in Nigerian deposit money banks (DMBs) between 2015 and 2024, combining a structured literature review of thirty empirical studies with a case-based analysis of seven purposively selected Tier-1 banks, Access Holdings, GTCO, Zenith Bank, UBA, Stanbic IBTC, FCMB and Wema Bank, for which technology-investment and fraud-loss figures are individually disclosed in audited financial statements for 2023-2024. Correlation and regression analyses were run on the four banks with matched data, alongside trend analysis of sector-wide fraud data from NDIC, NIBSS and FITC. The findings show that AI-powered audit tools, particularly data mining and image recognition, show a significant positive association with audit quality in several of the reviewed studies. In contrast, machine learning applications show a weaker, less consistent link at the firm level. The bank-level analysis found a negative, though not statistically significant given the small sample, relationship between technology investment and the percentage change in fraud loss (Pearson r = -0.39, n = 4). In contrast, an exploratory sensitivity analysis excluding Zenith Bank produced a much stronger correlation (r = -0.97, n = 3), a result treated as indicative only given the three observations. The evidence suggests that heavier technology investment is associated with smaller increases in fraud loss in some banks, but not universally. The study concludes that AI-powered audit tools hold real promise for Nigerian DMBs but remain constrained by data quality, infrastructure, skills, timing lags, and regulatory gaps. Recommendations are made for deliberate investment in data infrastructure, staff training, and phased regulatory guidance.