Internal Audit and Artificial Intelligence: A Narrative Review of Governance Conditions for Audit Effectiveness

Abstract

Artificial intelligence (AI) is reshaping organizational governance and, within it, the internal audit function. Research on AI in auditing has focused largely on external audit, on specific technologies or on ethical principles, and few contributions connect the theoretical foundations of governance and internal audit with the governance challenges that AI introduces. This paper therefore asks how AI is transforming internal audit and which governance conditions are required for this transformation to strengthen, rather than undermine, audit effectiveness, understood as analytical coverage, detection capacity, independence and quality of reporting. Based on a narrative and integrative review of 66 sources published between 1932 and 2026, and drawing on agency, contingency, neo-institutional, stakeholder and sociomaterial perspectives, the paper distinguishes data analytics, automation and AI in the strict sense, weighs the benefits of AI against its costs and limits, and develops a conceptual framework linking AI capabilities, AI governance mechanisms, the auditor's role and audit effectiveness, from which five theoretical propositions are derived. The review suggests that AI may enhance audit efficiency, broaden analytical coverage and strengthen fraud and anomaly detection, while simultaneously introducing governance risks related to algorithmic opacity, data quality, automation bias, accountability and auditor independence. The paper contributes by treating AI both as an audit tool and as an audit object, and by differentiating the managerial and regulatory implications for internal audit leaders, audit committees, senior management and regulators. It outlines precise research questions, hypotheses and designs for future empirical work, particularly in resource-constrained and public-sector contexts such as Moroccan public establishments.

Classification JEL : M42, M15, O33.

Paper type: Narrative literature review with a conceptual aim.