Autonomous AI in audit and control: Toward a governance framework for algorithmic assurance

This study develops a conceptual governance framework for integrating autonomous artificial intelligence (AI) into audit and control systems. It addresses the challenge of embedding AI as an active decision actor while preserving assurance integrity, accountability, and governance coherence. Using a systematic conceptual review, the study synthesizes insights from Agency Theory, Sociomateriality, Accountability Theory, and Institutional Theory. Key constructs, including algorithmic delegation, hybridized judgment, accountability displacement, and legitimacy pressures, are identified and organized into a layered framework spanning algorithmic, human, and institutional dimensions. The framework suggests that the adoption of AI can positively impact audit effectiveness and risk identification but creates hybridized judgment replacement and accountability. Symbolic adoption can occur due to institutional pressures restricting improvements in substantive assurances. Proper governance alignment at human, algorithmic and institutional levels can help to address these risks, and hold accountability, ethical control, and the quality of audit. The framework can help regulators, auditors, and organizations to structure the adoption of AI, define roles and responsibilities, and establish governance structures that ensure transparency and integrity of assurance in AI-facilitated audit environments. It provides a rigorous basis of empirical study and practical application in AI mediated audit systems.