Artificial Intelligence and Earnings Management Detection Among Listed Firms in Nigeria

This study examines the role of artificial intelligence (AI) in detecting earnings management among 151 listed firms on the Nigerian Exchange Group (NGX) over the period 2020–2024. Earnings management remains a persistent challenge in emerging economies like Nigeria, where regulatory enforcement gaps, weak institutional frameworks, and information asymmetry create fertile ground for managerial discretion in financial reporting. Traditional detection models—particularly accrual-based approaches rooted in the Modified Jones Model—have demonstrated well-documented limitations in capturing the increasingly sophisticated patterns of earnings manipulation. This paper argues that AI-driven analytical techniques, encompassing machine learning algorithms and natural language processing, offer a substantively more robust and adaptive toolkit for identifying both accrual-based and real earnings management. Employing an ex-post facto research design and panel regression analysis, the study models earnings management detection as a function of AI adoption indicators while controlling for firm size, industry type, leverage, profitability, growth opportunities, auditor type, board size, and board independence. Secondary data were sourced from audited annual reports, the Nigerian Exchange Group factbook, and the Securities and Exchange Commission database. The findings reveal that AI-based detection mechanisms significantly enhance the identification of earnings management practices, and that this relationship holds after accounting for firm-level and governance characteristics. The study contributes to knowledge by providing empirical evidence from an emerging market context, bridging the gap between AI adoption literature and earnings quality research in sub-Saharan Africa. Policy implications for regulators, auditors, and corporate governance practitioners are discussed.