Workers vs. The Algorithm: Simulating Collective Action in Gig-Economy Platforms

Kristina Lewandowska

Keywords: Algorithmic Collective Action, Gig Economy, Agent-Based Modeling, Labor Withholding, Multi-Agent Systems, Platform Governance, Simulation

TL;DR: We develop an agent-based model to simulate gig worker collective action against platform algorithms, revealing how network structure, worker heterogeneity, and algorithmic responsiveness determine the success of coordination

Abstract: Algorithmic Collective Action (ACA) represents a emerging form of coordination where participants in algorithmically-mediated systems organize to improve their outcomes. This paper presents a novel agent-based model to simulate ACA in gig-economy platforms, specifically modeling labor withholding strategies like DeclineNow. Our simulation framework captures the dynamic interplay between a heterogeneous population of workers, their communication networks, and an adaptive platform pricing algorithm. Through extensive experiments on our GigACA-Sim benchmark, we analyze the impact of network topology, platform responsiveness, and worker heterogeneity on the formation of critical mass and the long-term sustainability of collective action. Results demonstrate that our model provides a more nuanced and realistic understanding of ACA dynamics compared to mean-field or static-game theoretic baselines, revealing key strategic insights for both workers and platform designers.

Submission Number: 31