Leadership governance under algorithmic mediation: reconceptualizing transformative leadership through an integrative review

Purpose This paper explores the evolving nature of leadership in the context of AI integration, focusing on the transformation of leadership roles, ethical challenges and the redistribution of power and decision-making authority. It aims to offer a conceptual framework that guides leadership theory and practice in the digital age. Design/methodology/approach Using a structured literature review of sources published between 2014 and 2024, the paper synthesizes insights from transformational leadership, ethical leadership and the Cynefin framework. It also examines key organizational and human-centric challenges associated with AI adoption, including resistance to change, ethical risks and competency development. Findings The review indicates that in AI-mediated systems, leadership is enacted primarily as leadership governance. This involves decision-rights allocation, accountability, oversight and auditability, and escalation and human override. It also includes contesting algorithmic outputs when needed. The three domains identified (human-centred, strategic and technical) function as resources for enacting governance. Legitimacy and identity pressures rise when leaders are held accountable for AI-influenced outcomes without equivalent interpretive access or discretionary control. Practical implications For leadership educators, the findings suggest the need to move beyond generic “AI literacy” toward governance-informed learning designs that develop judgement under algorithmic mediation. Curriculum modules can explicitly combine ethical reasoning, identity work, and decision-making under algorithmic uncertainty. Originality/value The paper contributes to leadership theory by reconceptualizing transformative leadership in AI-integrated contexts as leadership governance, a sociotechnical condition defined by decision-rights allocation, institutional accountability, oversight, and enforceable escalation and override.