Artificial intelligence is increasingly transforming enterprise management systems through AI-enabled scheduling, productivity analytics, performance monitoring, compliance triaging, automated recommendations, and decision-support tools. While these technologies offer scalability, consistency, and operational efficiency, they also introduce serious governance concerns related to transparency, accountability, autonomy, contestability, procedural fairness, auditability, and organizational trust. Existing algorithmic management literature has primarily examined surveillance, worker control, autonomy erosion, and work intensification, with comparatively limited attention to governance architectures that preserve meaningful human oversight in AI-enabled enterprise work systems. This paper addresses that gap by proposing Human-in-the-Loop AI Management (HITL-AM) as a sociotechnical framework for responsible algorithmic governance. The framework conceptualizes AI-enabled management systems as governed human-AI work systems in which algorithms augment managerial decisions while human actors retain accountability, contextual judgment, escalation authority, ethical interpretation, and final responsibility for consequential outcomes. The paper introduces five governance dimensions: human oversight, explainability, contestability, autonomy preservation, and auditability. It also develops a governance maturity model, enterprise application scenarios, research propositions, and implementation indicators. The paper contributes to work design, responsible AI, algorithmic management, and enterprise governance literature by shifting the discussion from whether humans should remain involved in AI-mediated management to how meaningful human involvement can be operationalized, governed, measured, and audited.
Ajaya Singh (Mon,) studied this question.