An Introduction to the Theory of Artificial Intelligence Governance in Companies

Abstract

Rapid advances in artificial intelligence are challenging the traditional structure of corporate administration by raising fundamental questions concerning decision-making authority, managerial competence, responsibility, and supervision. The central issue is no longer limited to the use of artificial intelligence as a tool for facilitating corporate activities, but extends to the possibility of transferring part of managerial authority to intelligent systems and recognizing them as effective elements within the structure of corporate governance. This study adopts a descriptive-analytical method with a forward-looking approach to formulate the theory of artificial intelligence governance in companies. Within this framework, a distinction is made among assistive, augmented, and autonomous artificial intelligence, while the proposed theory focuses on the level at which an intelligent system moves beyond a supporting role and participates independently in corporate decision-making and management. Two principal institutional models are conceivable: replacing the board of directors with an artificial intelligence system, or establishing an independent AI governance layer alongside existing corporate organs. The legal recognition of such models raises significant challenges, including the absence of legal personality for artificial intelligence, uncertainty regarding the attribution of liability, the need to define the scope of AI authority, and the requirement for effective human oversight. The implementation of AI governance also depends on principles such as transparency, data governance, continuous auditing, accountability, security, AI literacy, and ongoing legal monitoring. The study concludes that full recognition of this theory is not presently compatible with the existing framework of Iranian company law; however, a gradual legal approach based on a special regulatory framework, limited recognition of the legal capacity of intelligent systems, and experimental implementation in small and low-risk companies could provide the basis for the emergence of a new form of corporate organization referred to as the “smart company.”