Autonomous data governance in Artificial Intelligence ecosystems: architectures, policies, and control mechanisms
Abstract
The expansion of artificial intelligence ecosystems has intensified the need for governance mechanisms capable of managing institutional responsibilities, access policies, regulatory obligations, and control procedures throughout the data lifecycle. In this context, the increasing incorporation of automated mechanisms into organizational processes creates opportunities for operationalizing governance policies while raising questions concerning decision legitimacy, administrative transparency, and human oversight. This study aims to analyze the technical, organizational, and normative foundations of autonomous data governance in artificial intelligence ecosystems, considering computational architectures, policy enforcement instruments, and accountability mechanisms required for implementation. The investigation is characterized as applied research with a qualitative approach and exploratory-descriptive objectives, grounded in bibliographic and documentary analysis of verifiable scientific publications and normative instruments. The theoretical foundation encompasses the National Institute of Standards and Technology's Artificial Intelligence Risk Management Framework, ISO/IEC 42001:2023, and Regulation (EU) 2024/1689, examined alongside academic contributions concerning organizational AI governance. The conceptual analysis distinguishes the automation of administrative procedures from the transfer of decision-making authority to computational systems. As a theoretical contribution, the study outlines a supervised autonomous governance model based on policy formalization, allocation of responsibilities, controlled decision execution, and preservation of auditable evidence. It concludes that operational autonomy must remain subject to institutional accountability and proportionate oversight mechanisms, while claims concerning efficiency or compliance improvements require specific empirical validation.