Algorithmic management is Digital Taylorism reborn: data-driven surveillance and automated decisioning reproduce scientific-management control in modern workplaces, so firms should optimize — not simply maximize — worker autonomy to balance efficiency and dignity.

Main Finding

Digital Taylorism is best understood as a historical continuation of Taylorist principles rather than a wholesale break: algorithmic management, real-time data collection, and pervasive digital surveillance reproduce the core logic of scientific management while creating new constraints on worker autonomy and dignity. The paper proposes "Optimum Labour Autonomy" — a theoretical point at which organizational coordination and employee discretion are balanced to maximize sustainable performance, worker well‑being, and meaningful work.

Key Points

• Historical continuity: Taylorism's emphasis on measurement, standardization, and managerial control persists through successive industrial revolutions and reappears in digital form under Industry 4.0.
• Digital mechanisms: Algorithmic management, workforce analytics, and continuous digital surveillance operationalize control in ways that echo time‑and‑motion studies but at much larger scale and greater informational precision.
• Dual orientations of autonomy:
  • Instrumental orientation: autonomy as a managerial lever to increase motivation, adaptability, and productivity (Human Relations, Socio‑Technical Systems, Job Characteristics, HRM).
  • Emancipatory orientation: autonomy as worker freedom, dignity, and control over work processes (Marxian/Labour Process Theory, Capability Approach, Self‑Determination Theory).
• Optimum Labour Autonomy: the paper argues against "more autonomy is always better" — instead it conceptualizes an optimal balance where sufficient discretion enhances intrinsic motivation and innovation without undermining coordination or organizational goals.
• Gig/platform work: platform-mediated labour exemplifies the paradox — flexibility combined with algorithmic direction and asymmetrical informational power (workers gain schedule flexibility but lose substantive decision rights and bargaining power).
• Theoretical synthesis: integrates literature from Taylorism, Labour Process Theory, Foucauldian surveillance, algorithmic management, socio‑technical and psychological theories into a unified analytical framework.
• Policy angle: calls for human‑centred governance, regulatory attention to algorithmic workplace governance, and labour policy that protects autonomy and dignity.
• Limitations acknowledged: primarily qualitative, secondary‑data systematic review; limited quantitative evidence and limited coverage of empirical cases from emerging economies.

Data & Methods

• Approach: interpretivist, qualitative, historical‑comparative analysis.
• Method: systematic literature review with thematic analysis and historical comparison.
• Sources: peer‑reviewed articles, books, conference proceedings, policy reports, and institutional publications (databases: Scopus, Web of Science, JSTOR, ScienceDirect, SpringerLink, Emerald, Taylor & Francis, Wiley, Google Scholar; reports from ILO, OECD, WEF, EU).
• Search strategy: Boolean queries combining terms such as "Taylorism", "Digital Taylorism", "Industry 4.0", "Algorithmic Management", "Labour Autonomy", "Gig Economy", etc.; backward citation tracing.
• Inclusion/exclusion: emphasis on influential theoretical works and peer‑reviewed empirical studies; excluded opinion pieces lacking empirical/theoretical support.
• Analysis: thematic coding of recurring concepts (managerial control, autonomy, surveillance, algorithmic governance) and comparison across historical periods; triangulation across disciplines.
• Empirical scope: no original primary data collection; conclusions drawn from synthesis of existing literature.

Implications for AI Economics

• Measurement & Incentives
  • Algorithmic monitoring dramatically reduces information asymmetries for firms (better measurement of effort, quality, time use). This alters incentive design and monitoring costs in principal–agent models — enabling tighter performance contracts but raising moral hazard and effort extraction concerns.
  • Economics of incentives should account for non‑pecuniary harms (intrinsic motivation, autonomy) that algorithmic control can erode; models that treat utility solely as paid effort may mispredict long‑run productivity and turnover.
• Labour market structure & rents
  • Platform firms and algorithmic intermediaries can capture informational rents and shift bargaining power away from workers. This has implications for wage-setting, monopsony power, and distributional outcomes in general equilibrium.
  • Skill complementarity: AI substitutes for routine tasks while complementing non‑routine cognitive skills — continuing wage polarization. Algorithmic governance may also compress wage variability by standardizing performance metrics.
• Firm organization & productivity
  • Optimal allocation of decision rights between algorithms, managers, and workers is an economic design problem: too little autonomy can lower innovation and intrinsic effort; too much can reduce coordination and scale economies. The "Optimum Labour Autonomy" concept suggests firms face a trade‑off that can be formalized and estimated.
  • Empirical productivity gains from automation may be overstated if costs to wellbeing, learning, or long‑term human capital accumulation are ignored.
• Policy and regulation
  • Regulatory interventions (transparency of algorithms, right to explanation, limits on surveillance intensity, collective bargaining rights for algorithmically managed workers) can affect firm incentives and equilibrium employment outcomes — potentially correcting market failures from informational asymmetries and power imbalances.
  • Antitrust and labour policy should consider algorithms as governance tools that shape market power, not just as production technologies.
• Welfare & measurement
  • Standard productivity metrics miss welfare dimensions (autonomy, meaningful work). AI economics should broaden evaluation frameworks to include psychological and capability measures when assessing technology impacts.
• Research agenda for AI economics
  • Quantify the trade‑off between monitoring intensity and productivity/worker welfare using firm‑level data (matching algorithmic monitoring logs with outcomes).
  • Structural models: incorporate endogenous autonomy choice into firms' optimization (principal–agent models where monitoring technology is a choice variable with costs and worker utility responses).
  • Field experiments: test alternative algorithm designs that preserve aspects of autonomy (e.g., human‑in‑the‑loop overrides, transparent performance feeds) to measure effects on output, error rates, turnover, and subjective well‑being.
  • Distributional studies: estimate how algorithmic governance affects wage inequality, bargaining power, and labor shares across sectors and countries (including emerging economies).
  • Policy evaluation: model and empirically test implications of transparency, data‑use restrictions, and collective rights on firm behavior and worker outcomes.
• Design recommendations for AI systems (economics‑informed)
  • Design algorithmic management systems that internalize autonomy costs (e.g., allow worker discretion where it yields learning or higher intrinsic motivation).
  • Incorporate feedback loops that monitor not only output but also indicators of worker strain, turnover risk, and error patterns — use these as constraints in optimization.
  • Experiment with hybrid governance: combine algorithmic coordination for routine standardization with delegated human decision rights for complex, non‑standard tasks.

Overall, for scholars and policymakers in AI economics, the paper highlights that algorithmic tools are governance institutions with distributional and incentive consequences. Treating them as neutral productivity enhancers misses critical trade‑offs between efficiency and autonomy; these trade‑offs should be central in both theoretical models and empirical analyses.