Beyond AI Ethics Washing: Toward AI Governance Guided by Organizational Virtue

As artificial intelligence becomes increasingly embedded in social life, organizations respond to its risks through ethical principles and governance arrangements, yet these commitments do not necessarily shape actual decisions. AI ethics washing may arise when ethical principles, governance commitments, and symbolic measures are used to project a responsible image while obscuring deficiencies in actual governance, calling for more effective governance. Drawing on conceptual clarification and normative analysis, this paper examines the organizational mechanisms and ethical consequences of ethics washing, and on this basis develops an organizational virtue governance framework to address the problem. It argues that the instrumentalization of ethical goals, the separation of ethics institutions from decision-making authority, the disconnect between procedures and ongoing practice, and the diffusion of responsibility across the division of labor may operate separately or in combination to produce AI ethics washing and may lead to erosion of trust, misplacement of responsibility, and ethics bashing. This organizational virtue governance framework comprises purposive good, institutional good, and procedural good. Purposive good requires ethical reasons to guide and constrain organizational goals; institutional good requires that those responsible for ethical judgment have commensurate access to information, resources, and authority; procedural good requires that reflection, participation, revision, and responsibility be sustained throughout the AI lifecycle.