Understanding Gig Workers Resistance to Algorithmic Control

Companies are increasingly using digital technologies and algorithms to control employees. Control measures that were originally performed by human managers are in part completely automated by technologies. Such systems require large amounts of behavioral data to enable constant surveillance and performance evaluation. Now, the use of algorithms for management and control triggers various forms of resistance behavior. Thus, the goal of the project is to gain a nuanced understanding of the mechanisms behind gig workers` resistance to algorithmic control. The research project focuses on the so-called gig economy, i.e. that part of the labor market in which temporary jobs are assigned flexibly on an ad-hoc basis. Well-known examples are so-called ride hailing services, such as Uber and Lyft, which control the behavior of their drivers through the mandatory use of an app. For example, not only is the route specified by the app, but the algorithm also tries to manipulate the drivers into accepting certain rides by means of incentives and sanctions. But also in the field of crowd working, algorithmic management is increasingly used. A nuanced understanding of the mechanisms behind gig workers` resistance to algorithmic control is urgently needed for research in this area, such as the socio-emotional and economic impact on workers, as well as ethical considerations for the use of algorithmic management. In addition, the results of this project also have important practical implications, whether for explaining the causes of negative consequences of algorithmic control, such as high turnover rates, or for designing algorithmic management to help improve worker job satisfaction.

Gig work platforms often promise autonomy and flexibility, yet this research shows that algorithmic management tightly controls gig work, shaping pay, task allocation, and performance evaluation in ways that are often opaque to workers. While these systems can create stress and uncertainty, they also offer opportunities for resilience and resistance. This project provides evidence-based insights and practical tools that go beyond describing resistance, offering specific strategies for fairer and healthier digital labor. A key finding is that gig workers respond very differently to algorithmic control. Some become trapped in cycles of stress, disengagement, and withdrawal, while others develop positive cycles of resilience, agency, and empowerment. How workers navigate these pressures depends largely on the strategies they develop. By cultivating coping strategies, such as seeking social support, and practicing proactive forms of resistance, workers can move from reactive, isolating responses to self-directed, empowering approaches. Platforms may play a supporting role, for example through transparent communication, peer support, or accessible training. The research thus provides a broader picture of resistance in the gig economy, taking into account not only open resistance but also individual, often covert coping strategies, and shows why traditional interventions such as transparency initiatives often remain ineffective. Another major contribution is the creation of a validated instrument to measure worker resistance, or "algoactivism." Previous studies described how gig workers resist or adapt to algorithmic pressures, but no systematic tool existed to capture these behaviors. This project developed a survey instrument, informed by expert interviews and gig workers, using text-mining to capture multiple dimensions of resistance. The scale allows researchers, policymakers, organizations, and workers themselves to measure, compare, and monitor forms of algoactivism and evaluate interventions across the gig economy. It also served to empirically validate the core research model and theoretical assumptions. The study also highlights the potential of modern large language models (LLMs) for studying digital labor. LLMs can accurately classify small forum datasets and efficiently annotate larger ones, producing high-quality data for transformer models like DeBERTa. Online discussions among gig workers provide rich temporal insights, yet their volume makes manual analysis difficult. Using LLMs, researchers can track behavioral changes over time, reduce coder bias, and increase reliability. This approach revealed shifts in discussion content, uncovering patterns that provide actionable insights into platform dynamics and worker experiences. Overall, these findings show that algorithmic management is not only technical but deeply socio-technical. Platforms and workers together shape gig work, but workers' agency and coping strategies are central to pushing back against harmful cycles, fostering empowerment, and negotiating their relationships with platforms. By integrating insights from coping, validated measurement tools, and computational methods, this research provides theoretical and practical innovations that support fairer, healthier, and more sustainable digital labor.