Research on algorithmic management has moved from isolated technical studies to a broad socio‑ethical critique driven by generative AI and regulation, identifying four key tensions—efficiency versus well‑being, control versus resistance, opacity versus governance, and platform-to‑corporate transitions—and urging multi‑stakeholder governance and targeted research.

Main Finding

The paper provides a mixed-methods bibliometric and qualitative synthesis of 340 peer‑reviewed articles (2017–2025) on algorithmic management. It documents a rapid maturation of the field — from early technical/coordination studies to a systemic socio‑ethical critique driven by COVID‑era diffusion and the rise of generative AI — and distills four core socio‑technical tensions that act as real‑world implementation barriers. The authors derive an actionable, multi‑stakeholder research and governance agenda to guide future work and policy.

Key Points

• Scope and growth

  • Sample: 340 English peer‑reviewed journal articles drawn from Scopus (search terms: “algorithmic management” OR “algorithmic work”), covering formative literature through end of 2025.
  • Field expansion: ~78.7% annual growth; publication surge after 2020, large spikes in 2024–25 linked to pandemic effects and generative AI capabilities.
  • Interdisciplinary spread: organizational behavior, information systems, HRM, sociology, and law; top outlets include European Labour Law Journal, New Technology, Work and Employment, and ACM HCI venues.

• Methodological approach

  • Blended review: integrates phenomenon‑focused (real‑world operational and ethical barriers) and field‑focused (disciplinary framings) strands.
  • Bibliometrics: performance analysis using Normalized Total Citations (NTC) and science mapping (Bibliometrix/Biblioshiny); thematic evolution compared across 2017–2022 vs. 2023–2025.
  • Qualitative synthesis: purposive sampling of top NTC articles until saturation (N=32); two independent coders (Cohen’s κ > 0.80); three‑stage coding from first‑order phenomena to four aggregate socio‑technical tensions.

• Four core socio‑technical tensions (implementation barriers)

• Coordination efficiency vs. psychological well‑being (algorithmic optimization drives productivity but erodes autonomy, increases technostress and precarization).
• Algorithmic control vs. worker resistance (informal “algoactivism,” account‑sharing, workarounds, and formal collective bargaining responses).
• Algorithmic opacity vs. structural governance (black‑box decisions create accountability gaps and demand for audits/transparency).

• Transition from platform labor to mainstream HRM (algorithmic practices are migrating from gig platforms into conventional organizations, raising new regulatory and managerial questions).

• Outcomes and contributions

  • Phase model of intellectual maturation: early technical matching → expanded control & surveillance → socio‑ethical and regulatory focus, accelerated by generative AI.
  • Actionable recommendations: multi‑stakeholder governance tools (audits, algorithmic impact assessments, worker‑facing dashboards), research priorities, and policy interventions.

Data & Methods

• Data collection

  • Source: Scopus query (“algorithmic management” OR “algorithmic work”) with start point anchored around 2015; dataset extracted in early 2026 and narrowed to peer‑reviewed English journal articles.
  • Screening: multi‑stage filtering (exclude conference papers, book chapters, non‑peer outputs), manual abstract screening to remove false positives; final N=340 articles.

• Quantitative analysis

  • Performance metrics: Normalized Total Citations (NTC) to rank influential works while correcting for publication age.
  • Science mapping: R Bibliometrix/Biblioshiny used to produce strategic thematic maps; centrality and density metrics to classify themes and track shifts across two time windows (2017–2022; 2023–2025).

• Qualitative coding

  • Sample selection: iterative sampling of highest‑NTC articles until theoretical saturation; final qualitative N=32.
  • Coding protocol: two coders trained on a subset (8 articles), achieved Cohen’s κ > 0.80; first‑order inductive codes (phenomena) → second‑order disciplinary categories → aggregate themes (four tensions).
  • Integration: blended phenomenon + field lens used to surface disciplinary blindspots and synthesize cross‑cutting tensions.

Implications for AI Economics

• Research directions and priorities

  • Model the trade‑offs: formalize the productivity vs. welfare trade‑off created by algorithmic management (e.g., models that map algorithmic control intensity to short‑run output gains and medium/long‑run worker well‑being/productivity).
  • Causal inference: prioritize causal empirical designs (field experiments, difference‑in‑differences, regression discontinuity, instrumental variables) using platform and HR administrative data to estimate effects on wages, hours, turnover, health, and firm performance.
  • Distributional analysis: quantify heterogeneity across worker types (platform vs. salaried; skill levels; bargaining power), sectors, and countries to assess impacts on inequality and labor market segmentation.
  • Behavioral & strategic responses: integrate worker resistance (algoactivism) into equilibrium models — dynamic effort choices, shirking/workarounds, account‑sharing, and unionization as endogenous responses to algorithmic rules.
  • Information asymmetry & market design: study how algorithmic opacity alters information flows between firms, workers, and regulators; analyze mechanisms (e.g., transparency dashboards, third‑party audits) as remedies in markets with asymmetric information.
  • Macro/labor market modeling: incorporate widespread adoption of algorithmic management into macro labor supply/demand and matching models to project sectoral employment shifts and reallocation costs.

• Methodological recommendations for economists

  • Data sources: platform logs, matched employer‑employee panels, administrative health/benefits data, surveys capturing perceived autonomy/technostress, and audit outputs.
  • Mixed methods: combine quantitative causal estimation with qualitative process tracing to capture institutional and behavioral mechanisms (e.g., how design choices trigger resistance).
  • Structural & computational approaches: use agent‑based models for diffusion and emergent collective action; structural estimation to evaluate welfare effects and optimal regulation under informational constraints.

• Policy and market design implications

  • Regulation design: evidence is needed on the effectiveness and cost of governance interventions (algorithmic impact assessments, mandatory transparency, worker participation rights) to design economically efficient safeguards that minimize deadweight loss while protecting welfare.
  • Contracting and incentives: study how contracts and compensation schemes should adapt when managerial decisions are delegated to algorithms (optimal incentive contracts, bonus/penalty design in the presence of automated monitoring).
  • Collective bargaining & institutions: analyze the role of unions and labor institutions in internalizing negative externalities and in negotiating algorithmic governance — important for models of bargaining power and wage setting.
  • Innovation vs. protection trade‑off: economists should quantify potential innovation slowdowns or compliance costs from stricter governance and compare against benefits (reduced stress, lower turnover, improved fairness) to inform balanced policy.

• Practical takeaways for empirical economists

  • Expect endogenous adoption: firms adopt algorithmic management as a response to competition, scale, and technological opportunities; empirical strategies must address selection and timing.
  • Measure intensity, not just presence: build indices for control intensity (monitoring frequency, decision automation share, opacity measures) to capture dose‑response effects.
  • Consider spillovers: algorithmic practices in one firm/sector can diffuse and alter labor market norms; account for spillovers in identification strategies.

Concluding note: the paper supplies a roadmap linking disciplinary debates to concrete research questions and policy levers. For AI economists, it signals rich, high‑priority opportunities for causal and structural work on productivity‑welfare trade‑offs, regulatory design, and the labor market consequences of algorithmic governance.