Gig platforms exercise managerial control through algorithms but dodge employer obligations, entrenching precarious, dependent work; the paper urges labour law reform—recognising dependent contractors, imposing platform liability for welfare, and mandating algorithmic transparency—to restore accountability.

Main Finding

Current labour-law categories and corporate rules fail to capture how platforms exert employer-like control through algorithms while avoiding employer-like liabilities. The paper argues that this mismatch produces structural dependence, algorithmic surveillance, and enforcement practices that—while not slavery in a property sense—constitute a form of "digital slavery" (economic bondage, coerced behaviour, lack of bargaining power and social protection). It calls for reconstructed labour law grounded in economic dependency (e.g., dependent-contractor status), platform liability for worker welfare, algorithmic transparency, social-security obligations, and specialised grievance/appeal mechanisms.

Key Points

• Algorithmic management replaces visible human supervision: platforms allocate work, set dynamic prices, rate performance, impose automated penalties, and deactivate workers via algorithms. Functional control exists even when contracts label workers as independent.
• Misclassification and the legal fiction of “platform neutrality”: standard-form contracts and corporate structuring enable platforms to claim they are neutral intermediaries, shifting risk to workers and evading statutory employer obligations and vicarious liability.
• Conceptual reframing: traditional tests (control, integration, economic reality) are weakened by technologically mediated control. The paper emphasizes a "dependent contractor" category to reflect economic dependence plus algorithmic control.
• Normative framing—“digital slavery”: not literal chattel slavery, but a condition created by economic dependence, surveillance, automated discipline, and lack of recourse (deactivation, no bargaining, absent social protection).
• Comparative legal review:
  • United Kingdom: intermediate “worker” category (Uber BV v Aslam) gives some protections but falls short of full employer liability.
  • United States (California): AB5/ABC test tightened classification; Prop 22 created a carve-out for app-based drivers (partial benefits, not employees), illustrating political pushback and regulatory discontinuity.
  • European Union: Platform Work Directive proposal establishes a presumption of employment when indicators of control (including algorithmic management) exist and pushes for algorithmic transparency.
  • India: Social Security Code 2020 recognises gig/platform workers as a category for welfare schemes but does not confer full employment rights—recognition without worker entitlements.
• Policy proposals advanced: legal recognition of dependent-contractor status; statutory platform obligations for social security and worker welfare; algorithmic transparency and contestability; specialised grievance mechanisms and clearer joint/joint-employer/vicarious liability doctrines.
• Central concern: without legal evolution, platforms will normalise exploitative, precarious labour under the rhetoric of innovation and flexibility.

Data & Methods

• Methodological approach: doctrinal and comparative legal analysis combined with normative/political-economy argumentation and synthesis of secondary literature.
• Sources and evidence:
  • Case law and legislation: Uber BV v Aslam (UK Supreme Court), California AB5 and Prop 22, EU Proposal for a Directive on Improving Working Conditions in Platform Work (COM(2021)762), India’s Code on Social Security (2020).
  • Academic and institutional literature: works by De Stefano, Prassl, Freedland & Kountouris, Srnicek; ILO reports on digital labour platforms.
  • Comparative policy review across jurisdictions to identify regulatory trends and gaps.
• Limitations (implicit in paper’s method): primarily conceptual and legal-analytical rather than original empirical fieldwork. The analysis relies on legal texts, court decisions, and secondary empirical findings (ILO, academic studies) instead of primary survey or administrative platform datasets.

Implications for AI Economics

• Power asymmetries driven by AI/algorithms: Algorithmic allocation, pricing, ranking and deactivation are economic instruments that create monopsony-like platform power, shifting risk and surplus extraction toward platforms. Economic models of platform labour must incorporate algorithmic governance as a channel of market power.
• Wage dynamics and distributional effects: Automated dynamic pricing and rating-driven access to work can depress effective wages, increase income volatility, and amplify inequality—affecting measures of labor market welfare, poverty risk, and social insurance needs.
• Labor supply elasticity & behavioural responses: The combination of nominal “flexibility” with algorithmic penalties likely alters worker reservation wages, multitasking incentives, and search behaviour—models should consider behavioral responses to reputational algorithms and deactivation risk.
• Regulatory interventions change platform incentives and equilibrium outcomes:
  • Reclassifying workers or imposing social-security obligations raises marginal labor costs; firms may pass costs to consumers, reduce supply of services, accelerate automation, or alter platform design (e.g., moving to fewer algorithmic controls).
  • Algorithmic transparency and contestability can reduce information asymmetries, enable audits, and potentially curb abusive automated discipline, but may also enable gaming or require confidentiality trade-offs.
  • Dependent-contractor status and joint-liability rules alter firms’ hiring, subcontracting and corporate-structuring strategies and can affect innovation incentives and scale economies.
• Measurement and empirical research priorities:
  • Quantify prevalence/intensity of algorithmic control (task allocation, dynamic pricing algorithms, deactivation rates) using platform data or field experiments.
  • Exploit natural experiments (e.g., AB5/Prop22, Uber v Aslam outcomes, EU directive adoption across member states) with difference-in-differences or regression discontinuity to estimate effects on wages, hours, platform supply, consumer prices, and automation adoption.
  • Audit algorithms and run algorithmic-impact assessments to evaluate fairness, wage impacts, and disemployment risk.
• Governance and design recommendations for AI in platforms:
  • Mandate algorithmic transparency, logging, and audit access for regulators and independent researchers.
  • Require contestability mechanisms enabling workers to appeal automated decisions and obtain human review.
  • Incorporate social-cost accounting for platform algorithms (internalize worker welfare impacts into platform optimization objectives).
• Broader macro implications: widespread platform regulation will affect labor-market structure, fiscal needs for social insurance (if platforms continue to avoid obligations), and the pace of AI-driven automation—key variables in forecasting employment, inequality, and public finance trajectories.

Suggested next research steps for AI-economics scholars: - Build causal estimates of algorithmic management’s effect on earnings and retention. - Model platform competition with regulatory constraints (e.g., payroll contributions, transparency costs) to predict equilibrium adjustments. - Design and evaluate policy counterfactuals (dependent-contractor regimes, universal portable benefits, algorithmic audits) using structural models and randomized/ quasi-experimental evaluations where feasible.

Summary: The paper links legal doctrine and normative critique to show that algorithmic control demands new legal categories and obligations. For AI economists, this means treating platform algorithms as central economic institutions shaping labor supply, bargaining power, wage formation, incentives for automation, and the effects of regulation.