ACCOUNTING AUDIT AND GENERATIVE ARTIFICIAL INTELLIGENCE: EVIDENCE, PROFESSIONAL JUDGMENT AND GOVERNANCE

Abstract

This theoretical essay examines how generative artificial intelligence affects accounting auditing, with emphasis on audit evidence, professional judgment, professional skepticism and governance of technology-assisted procedures. The study adopts a qualitative, bibliographic and documentary approach, supported by thematic analysis of Brazilian auditing literature, national auditing standards, data protection legislation and institutional documents on responsible AI use. Its objective is to discuss the conditions under which generative AI can support accounting audit activities without weakening independence, traceability, confidentiality and the auditor's technical responsibility. The discussion indicates that AI may assist document reading, organization of working papers, screening of accounting entries, preliminary contract analysis, control assessment and communication of audit findings. However, uncritical or undocumented use may intensify risks related to inaccurate outputs, data exposure, algorithmic opacity, operational dependence and insufficient audit evidence. The paper concludes that generative AI should be treated as an auxiliary analytical resource, subject to human validation, sufficient documentation and governance policies capable of preserving audit reliability.