The rapid diffusion of artificial intelligence (AI) in managerial decision making has revolutionized the use of performance for resource allocation and control by firm. But there are also critical questions about the ethics, accountability and governance of algorithmic systems. We suggest an integrated view of responsible AI in management which bridges performance-driven interest with ethics and accountability mechanisms. Bringing together insights from management theory, information system studies and organizational ethics, this article conceptualizes responsible AI as dynamic capability and part of managerial practices rather than a static technical feature. We derive a series of critical trade-offs between efficiency, fairness, transparency and managerial discretion from our conceptual model and consider how organizations may address these tensions via human-algorithm collaboration (HaC), explainability practices, and accountability mechanisms. The model emphasises the judge of manager to direct the decision-making process and it delegates responsibility through levels of organization. This paper makes several contributions by synthesizing various research streams, providing theoretical knowledge on the impact of AI on managerial decision-making and practical implications on how to design and manage a high-performing, ethically behaving AI system. The paper contributes to the existing literature on digital transformation, Corporate social responsibility and the future of management in algorithmically organized companies.