Shadow Accountability under Hybrid Algorithmic–Human Management: Consequences for Psychological Safety, Autonomy, and Middle Management

Abstract

Purpose: The paper addresses a gap in the literature: while research has examined algorithmic management, psychological safety, and middle-management discretion separately, little is known about how accountability is redistributed through governance artifacts under hybrid algorithmic–human arrangements. Aims: The primary aim is to theorize how accountability allocation, purpose signaling, and explainability influence psychological safety, autonomy, and discretion. A secondary aim is to identify status inferences and discretion boundaries as mediators linking governance to climate outcomes. Design/methodology/approach: A qualitative meta-synthesis of Q1/Q2 articles (2023–2025) was conducted across organizational behavior, HRM, information systems, and organization theory. Transparent inclusion criteria, PRISMA-style screening, and JBI appraisal produced a final corpus of 36 studies. Coding combined a priori constructs with inductive theme development. Findings: Four themes emerged: shadow accountability in dashboards and appeals; compression of middle-management discretion under metricized oversight; fragile psychological safety shaped through purpose and transparency; and status externalities that dampen voice. Recent field and experimental studies confirm heterogeneous productivity gains and displacement risks. Limitations of the study: Reliance on published research limits visibility into emerging contexts and precludes effect-size estimation. Practical implications: Role-anchored accountability maps, override rights, and recognition practices can preserve candor and trust. Originality/value: The framework extends algorithmic-management theory through governance design and provides actionable pathways for creating psychological safety.