From AI-Generated Audit Signals to Accountable Audit Action: A Traceable Skepticism Framework for AI-Assisted Auditing

Kobra Meskini¹ Amirhossein Panji Pazani²

1) Ph.D. Student in Accounting, Islamic Azad University, Tonekabon Branch
2) M.A. Student, Shafagh University, Tonekabon

IOI Code : ACEE-ACABDE
Publication : 5th International Conference on Researches in Accounting, Economics, Management & Humanities(meaconf.com)
Abstract :
    Artificial intelligence (AI) is increasingly being integrated into auditing to identify anomalies, analyze extensive datasets, and support the evaluation of complex accounting estimates. However, the production of an AI-generated signal does not automatically establish the existence of a misstatement or determine the appropriate audit response. Auditors must interpret algorithmic outputs, evaluate their evidential relevance, consider competing explanations, and retain responsibility for professional judgments. This conceptual paper develops a traceable skepticism framework connecting AI-generated audit signals to accountable audit action. Drawing on research on professional skepticism, auditor reliance on AI, technology-enabled audit evidence, and the ethical implications of algorithmic decision-making, the paper conceptualizes skepticism as a process that must be demonstrated through observable and documentable audit activities. The proposed framework comprises five interconnected stages: signal qualification, alternative-explanation generation, corroborative and contradictory evidence evaluation, proportionate audit response, and accountability documentation with controlled feedback. Six theoretical propositions explain how evidence traceability, calibrated reliance, auditor control, balanced evidence evaluation, and accountability mechanisms may influence skeptical audit action. The framework contributes to the conceptual discussion by integrating these elements into a single process-oriented model. It also provides practical implications for audit firms, engagement teams, and developers of AI-assisted audit systems. The framework remains a conceptual proposal requiring empirical evaluation.

Keywords : artificial intelligence auditing professional skepticism audit evidence auditor judgment accountability