The digital economy poses fundamental, systemic challenges to audit standards historically anchored in the industrial-era paradigm. Employing normative research methods, this paper identifies operational dilemmas across three dimensions-audit objects, audit evidence, and audit procedures-with particular attention to the systemic challenges introduced by generative artificial intelligence (AI) applications. Moving beyond surface-level symptoms, the analysis probes the underlying theoretical crisis: the systematic failure of foundational assumptions that have long sustained traditional auditing practices. The paper proposes a four-tier adaptive reconstruction pathway encompassing: (1) reinterpretation of core concepts, (2) refinement of theoretical frameworks, (3) transformation of standard-setting mechanisms, and (4) development of supporting competencies. Key innovations include a digital risk-oriented audit model, data lineage rules, a multi-dimensional materiality framework, and an agile standard-setting system. This study systematically delineates the multidimensional impact of generative AI on audit standards and constructs an integrated analytical framework linking operational dilemmas, theoretical challenges, and adaptive reconstruction. The central argument is that the adaptive reconstruction of audit standards for the digital economy transcends mere technical patching. It necessitates a paradigm shift-from physical-trace tracking to data-lineage verification, and from monetary-amount-centric judgment to multi-dimensional materiality assessment-while maintaining synergistic continuity with theoretical adjustment. By clarifying the boundaries and interconnections between paradigm shift and theoretical adjustment, and drawing on international comparisons, the research reveals China’s strategic opportunity to transition from passive adaptation to active leadership in global digital auditing standards.