Governance in the Age of AI Leadership: From Advisory Systems to Organizational Authority

The question of whether artificial intelligence can occupy the role of Chief Executive Officer has shifted from theoretical provocation to live experiment. In October 2025, Kazakhstan’s sovereign wealth fund Samruk-Kazyna formally elected SKAI — an AI system — as an independent board director with claimed voting rights, though the legal enforceability of that authority remains disputed under Kazakh law. Researchers at the Wharton Mack Institute and INSEAD’s Center for Corporate Governance have begun formally comparing AI and human boards on governance criteria. The Dataiku Global AI Confessions Report (Harris Poll, March 2025), surveying over 500 CEOs across the US, UK, France, and Germany, found that 94% believe AI could offer equal or better counsel than at least one of their current board members. This paper examines what these developments mean: what AI can do in an executive capacity, where it demonstrably cannot, and what governance architecture is required if organizations are to move from AI-assisted leadership to AI-principal leadership responsibly. Drawing on principal-agent theory (Jensen & Meckling, 1976), organizational legitimacy theory (Suchman, 1995), and institutional governance frameworks (Ostrom, 1990), it introduces a six-level AI Authority Maturity Model and proposes governance as a form of strategic infrastructure rather than a compliance cost.