A synthesis of 751 studies finds platform infrastructures and embedded algorithms reshape work processes and create uneven worker outcomes, prompting multi-level governance responses; the authors propose a mechanism-based sociotechnical framework and identify algorithmic fairness, worker well-being, and sustainable platform governance as priority research gaps.

Main Finding

The paper provides a systematic, bibliometric and integrative review of platform/gig-work research (751 articles, 2018–2026), identifying three evolutionary stages and six dominant research themes, and proposes a mechanism-based sociotechnical framework showing how digital technologies (platform infrastructures and algorithms) reconfigure labor processes, produce differentiated worker outcomes, and provoke governance responses under specific boundary conditions.

Key Points

• Dataset and scope: 751 articles (2018–2026) drawn from the Web of Science Core Collection and cross-checked with Scopus.
• Analytical approach: bibliometric mapping using CiteSpace (publication trends, collaboration networks, keyword co-occurrence, burst-term analysis), combined with integrative thematic synthesis.
• Evolutionary stages: the field evolves through three identifiable stages (mapping and tracing of temporal development in the literature).
• Six dominant thematic streams:
• Digital platforms as sociotechnical workplaces
• Algorithmic management
• Health and precarity
• Work experience
• Differentiated groups (heterogeneous worker outcomes)
• AI-driven governance
• Core conceptual contribution: a mechanism-based sociotechnical framework that links platform infrastructures and embedded technologies → reconfigured labor processes → heterogeneous worker outcomes → governance responses, conditioned by institutional, market, and worker-level boundary conditions.
• Research gaps & future directions flagged: algorithmic fairness, labor well-being, and sustainable platform governance.

Data & Methods

• Data: 751 peer-reviewed articles published 2018–2026; primary index: Web of Science Core Collection; validation: Scopus cross-check.
• Tools: CiteSpace for bibliometric/network analyses.
• Analyses conducted:
  • Longitudinal publication trends to trace field growth and stage transitions.
  • Collaboration structure mapping (author, institution, country networks).
  • Keyword co-occurrence and clustering to surface thematic concentrations.
  • Burst-term detection to identify emergent topics and turning points.
• Integration: bibliometric results synthesized with qualitative reading to build a mechanism-based sociotechnical framework and to delineate six major research streams.

Implications for AI Economics

• Labor supply and matching: algorithmic platforms reshape how workers discover and select tasks (search frictions, dynamic pricing), altering labor supply elasticity and matching efficiency—models should incorporate algorithmic interfaces and dynamic information frictions.
• Productivity and compensation: algorithmic management affects measured productivity, task allocation, and pay-setting (surge pricing, piece rates, micro-tasks); economists should quantify causal impacts on earnings, effort, and productivity using quasi-experimental and structural approaches.
• Distributional effects and heterogeneity: “differentiated groups” theme implies heterogenous treatment effects across demographics, skill levels, and geographies—necessary for welfare analysis and policy design addressing inequality.
• Algorithmic fairness and discrimination: algorithm-driven assignment and reputational systems can generate biased outcomes; economic models and empirical work should test for disparate impacts and design corrective mechanisms (e.g., counterfactual auditing, fairness-constrained optimization).
• Worker well-being and precarity externalities: health and precarity findings suggest non-wage costs (stress, injury, unstable income) that should be included in welfare calculations and cost–benefit analyses of platform regulation.
• Governance and policy design: evidence on AI-driven governance points to multi-level regulatory levers (platform-level algorithmic governance, labor law, data protection). Economic evaluation needed for trade-offs (innovation vs. worker protection) and for designing optimal incentives/regulations (e.g., liability rules, transparency mandates, collective bargaining constraints).
• Methodological suggestions for AI economics:
  • Combine large-scale platform data with randomized/quasi-experimental designs to estimate causal effects of algorithms on labor outcomes.
  • Use structural models to simulate policy interventions (taxes, minimum earnings, transparency rules) accounting for platform algorithm dynamics.
  • Incorporate interdisciplinary measurement (health, subjective well-being) to capture non-monetary outcomes.
  • Leverage bibliometric maps to identify underexplored empirical settings and cross-disciplinary methods.
• Research agenda pointers: quantify algorithmic management’s aggregate labor-market effects, assess long-run human capital implications of platform work, design algorithmic governance mechanisms that balance efficiency with equity, and develop metrics for algorithmic fairness and worker well-being suitable for policy evaluation.