Agentic AI and the ethics of leadership maintenance: rethinking responsibility in algorithmic organizations

Purpose This is a conceptual, theory-building paper that develops operational scaffolding in the form of testable propositions to redefine ethical AI maintenance as a strategic leadership responsibility rather than a technical or compliance task. Design/methodology/approach An interdisciplinary literature review develops operational scaffolding and testable propositions organized into a three-part leadership-centered framework: (1) post-deployment oversight, (2) ethical monitoring systems and (3) alignment with organizational and societal values, with practical indicators and SDG linkages. Findings Agentic AI systems operate across high-impact domains; however, few organizations assign leadership-level accountability to their ethical maintenance. Without executive ownership, post-deployment risks such as bias drift, opacity and ethical breakdowns remain unaddressed. Research limitations/implications This study lacks empirical testing as a conceptual paper. Future research should pilot the framework in diverse organizational contexts, evaluate leadership training interventions and develop validated auditing tools. Practical implications The framework provides actionable governance pathways and supports the integration of ethics into leadership development, ensuring that accountability is embedded at the executive level. Originality/value By shifting the AI ethics conversation from abstract principles to enforceable executive accountability, this study offers a model that bridges ethics and leadership practices. Ultimately, ethical AI maintenance is not a compliance exercise but a strategic leadership responsibility, transforming governance from principle into practice and offering a path toward trustworthy AI in high-stake contexts.