Algorithmic management is widespread across European workplaces and, when highly autonomous, opaque or rigid, correlates with organizational brittleness and worker anxiety; conversely, transparency and human-in-the-loop governance are linked to greater trust and adaptive capacity.

Main Finding

Algorithmic management (AM) is widespread across European workplaces—especially for scheduling and task allocation—and exhibits configurations that both create and can mitigate BANI (brittle, anxious, nonlinear, incomprehensible) vulnerabilities. High levels of algorithmic autonomy, opacity, and rigidity correlate with organizational brittleness, elevated employee anxiety, and reduced comprehensibility of decisions. Conversely, transparency-by-design, explainable AI, and human-in-the-loop governance are associated with higher trust, perceived fairness, and adaptive capacity, enabling more human-centric leadership.

Key Points

• Research questions / hypotheses
  • RQ1 / H1: Higher AM autonomy, opacity, and rigidity → more employee anxiety, less autonomy, greater brittleness.
  • RQ2 / H2: Transparency, explainability, and meaningful oversight mitigate BANI vulnerabilities and increase trust/adaptiveness.
  • RQ3 / H3: Responsible AM moderates the efficiency–leadership trade-off, supporting adaptive & human-centric leadership.
• Empirical prevalence (EU averages, AIM-WORK / OECD sources)
  • Automatic allocation of working time (rosters/shifts): ~24% of workers.
  • Automatic allocation of tasks: ~21%.
  • Automated rewards (points/ratings): ~13%; automated benchmarking/dashboards: ~12%.
  • Automated task instructions: ~10%; automatic cancellation of shifts: ~7%; automatic determination of work speed: ~5%; use of online customer ratings for control: ~4%.
  • ~30% of EU workers reported using an AI tool at work at least once in the prior 12 months (AIM-WORK ~29.9%).
• Cross-national and occupational heterogeneity
  • Higher exposure in Spain, Poland, Ireland, Romania; lower in Greece, Bulgaria, Hungary, Netherlands.
  • AM intensity and forms vary significantly by sector, occupation and national institutional context.
• Mechanisms and consequences
  • Rigid, opaque AM reduces worker discretion and sense-making → anxiety and brittleness (fragility under shocks).
  • Nonlinear interactions between humans and algorithms can produce disproportionate, unpredictable outcomes.
  • Responsible design (transparency, explainability, participatory design, human oversight) reduces negative psychological outcomes and enhances resilience.
• Original contribution
  • Empirically links AM configurations to BANI vulnerabilities using large-scale European data and integrates responsible AI governance with theories of human-centric/adaptive leadership.
• Limitations
  • Secondary, cross-sectional data limits causal inference and longitudinal insight.
  • Leadership outcomes are proxied by organizational/work-design indicators (autonomy, perceived fairness, oversight), not direct behavioral measures.

Data & Methods

• Mixed-methods study:
  • Systematic literature review (2020–2025) synthesizing conceptual and empirical research on AM and mitigation strategies.
  • Secondary quantitative analysis of large-scale European datasets:
    • AIM-WORK survey (De Cuyper et al.; Gonzalez Vazquez et al., 2025) — worker-reported AM practices and AI-tool usage.
    • OECD report “Algorithmic Management in the Workplace” (OECD, 2025) — cross-national statistics on adoption, design features, autonomy, transparency, human-in-loop.
  • Analytical approach: descriptive statistics and cross-tabulations to map prevalence/configurations and associations with BANI-related indicators; theoretical interpretation linking patterns to leadership dynamics.
• Operationalization notes:
  • BANI-related outcomes proxied by measures such as perceived autonomy, participatory practices, communication clarity, perceived fairness, and presence of human oversight.
• Suggested future methods by authors:
  • Longitudinal and qualitative designs to establish causality and to observe how responsible AM shapes leadership behaviors and organizational resilience over time.

Implications for AI Economics

• Trade-off between efficiency and resilience
  • Economic models should incorporate an efficiency–fragility trade-off: gains from algorithmic optimization can increase short-run productivity but also raise systemic fragility (brittleness) and downside risk. Treat human oversight/participatory governance as costly investments that enhance resilience (like insurance).
• Labor market and distributional effects
  • Heterogeneous exposure implies differential impacts on job quality, stress, and bargaining power across occupations and countries—potentially widening inequality between workers in algorithm-intense roles and others.
  • Heightened anxiety and reduced autonomy can affect labor supply (absenteeism, turnover) and productivity in ways not captured by simple effort-productivity mappings.
• Valuation of explainability & human-in-the-loop
  • Explainability, transparency, and oversight have economic value beyond compliance: they can raise trust, reduce turnover and error costs, and improve adaptiveness in the face of shocks. These should be incorporated into cost–benefit analyses and investment decisions for firms and policymakers.
• Policy and regulation
  • Cross-national variation suggests regulatory and institutional context matters for AM outcomes. Policy levers (transparency requirements, worker participation mandates, auditability, liability rules) can alter firms' incentives and the social returns to algorithmic adoption.
  • Regulators should weigh short-term productivity gains against systemic risk and worker welfare; targeted regulation for high-fragility domains (logistics, care, public services) may be warranted.
• Modeling recommendations for researchers
  • Include BANI-like state variables (fragility, anxiety, incomprehensibility) in organizational and macro models to capture nonlinearities and tipping points from algorithmic deployment.
  • Model endogenous design choices: firms choose level of automation, opacity, and human oversight balancing cost, efficiency, and expected resilience; allow for heterogeneous firm types and institutional constraints.
  • Empirically estimate dynamic effects (panel/experimental designs) to identify causal impacts of AM features on productivity, turnover, mental health, and shock response.
• Empirical priorities
  • Measure the monetary costs of brittleness (risk of failure, disruption recovery) and benefits of governance (reduced error, faster adaptation).
  • Sectoral and country-specific studies to quantify distributional impacts and to guide policy calibration.
  • Randomized or quasi-experimental evaluations of transparency/explainability/human-in-the-loop interventions to estimate their ROI.

Short summary: AM increases efficiency but can produce economic costs via fragility, anxiety, and nonlinear harms; responsible AM design and governance create measurable economic value by improving trust, fairness, and resilience, and should be incorporated into models, firm decisions, and policy design.