Skip to content

Surveillance and Resistance in Amazon’s Growing Platform Ecosystem

PUBLISHED

Sarrah Kassem (@KassemSarrah) is Lecturer and Research Associate in Political Economy at the Institute of Political Science at the University of Tübingen.

This post is part of a symposium on Worker Surveillance & Collective Resistance. Read the rest of the posts here.

***

The world of work is constantly in flux, evolving and co-evolving along with changes in societal, political-economic, and technological conditions. These changes not only come to structure working conditions through new forms of management and surveillance, but also inform the bounds of collective labor organization and resistance aimed at improving working conditions.

Perhaps nowhere is this observation more evident than in the platform economy. Via the technological infrastructure of the Internet, digital platforms mediate the sale and exchange of various products and services between different groups. These have come to constitute platform capitalism, which is situated in capitalism’s larger regime of accumulation. Platforms have emerged and developed at different moments, from Google and Amazon in the 1990s, to Facebook and Amazon Mechanical Turk (MTurk) in the mid-2000s, to Airbnb, Uber, and Deliveroo after the economic crisis of 2007-2008. As they continue to develop and some grow into monopolies, platforms have become—to differing degrees across the globe—an intrinsic part of our daily lives.

At the core of how platforms organize work and workers lies their systems of surveillance. Platforms differ markedly in how they instrumentalize technology to mediate and oversee the labor process—and this differentiation has implications for how workers resist such oversight. I examine these points in my larger analysis of how platforms alienate workers, and how workers organize in resistance to alienation, in my book Work and Alienation in the Platform Economy: Amazon and the Power of Organization. I argue that it is helpful to analyze the realities of platform workers along two dimensions: the nature of the platform (i.e., how platforms mediate employment relations, especially with regard to whether workers are location-based or web-based) and the nature of the work (i.e., how labor is remunerated, meaning by a traditional hourly wage or by “gig”).

To grasp the implications for the world of work(ers) more closely, I contrast two of Amazon’s platforms in its growing eco-system: (1) its e-commerce platform, where workers in the company’s location-based warehouses are paid a relatively fixed hourly wage, and (2) MTurk, a digital labor platform facilitating the online outsourcing of remote work on microtasks, which can be regarded more generally as constituting part of ‘ghost work.’ On the latter platform, remote workers based predominantly in India and the US are paid precariously when and if a submitted task is approved. By comparing these two platforms along two dimensions—the nature of the platform and the nature of the work—it is possible to understand the growing role of technology in the labor process, which allows for both old and new ways of surveilling and managing workers. These engulfing dimensions of surveillance in turn inform how workers resist employer practices in their own traditional and alternative ways.

Let’s first direct our attention to the Amazon warehouses. As these workers labor under Taylorist, factory-like conditions, they are assigned specific tasks, including prepping, stowing, picking and packing items. Workers encounter surveillance both through the social eye of their supervisors and managers, as well as digitally through the devices they use in the labor process (like computers or hand scanners). At the core of the labor process is the algorithmically managed regime of productivity, which requires fulfilling ‘Units Per Hour’ (UPH) rates. These differ based on the assigned task and order volume of the shift ahead. Thus, although workers are paid by the hour, they are evaluated based on their piecework. These various forms of surveillance structure the labor process and discipline workers, while alienating them from the very labor they carry out. In the process, both forms of surveillance also limit certain forms of subtle everyday resistance such as laboring at a slower pace, since workers may be labeled as ‘low performers’ and thereby risk their contract extension.

At the same time, the nature of the platform, which concentrates workers within warehouses, allows for communication and the formation of solidarity based on various class-based, gendered and racialized subjectivities. This embodied solidarity is especially crucial, given that workers are hired on different contracts (fixed, permanent, seasonal, subcontracted); face different material realities; and encounter Amazon’s union-undermining and union-busting tactics. Workers have been navigating these and organizing in various ways in their local and national contexts. Health and safety concerns and performance pressure—both of which are engendered and exacerbated by employer surveillance—are major factors driving workers to collectively organize. They may engage in walkouts, picket lines and strikes (where legally permissible), especially during peak season of Black Friday and Christmas, and campaign for unionization and collective bargaining agreements. These are possible given the nature of the work that typically categorizes these workers as Amazon employees. In short, the nature of this platform permits more traditional forms of resistance in the face of employer surveillance.

Let us now turn our attention to the starkly different case of MTurk. Workers on MTurk labor online in hyper-Taylorized digital production lines of ‘Human Intelligence Tasks’ (HITs). These HITs—posted by so-called ‘requesters,’ who can vary from individuals, starts-ups, corporation and universities—can include answering surveys, digitalization tasks, tracing 3D objects, or differentiating between images. As globally distributed workers participate in the production of data, often used for machine learning algorithms for AI, their labor is mediated exclusively through the interface. These workers are only paid precariously per gig, so income is neither guaranteed nor stable, and ranges from a few cents to several US Dollars. As ‘independent contractors,’ MTurk workers are left with no benefits, insurances or a guaranteed wage. The larger gig economy has become renowned precisely for this kind of precarious work.

Given the vulnerabilities of gig work and the web-based nature of the platform, technological surveillance is part and parcel of the labor process. Centered on algorithmic management, the interface measures the exact time needed for each task and the worker’s approval rating of those submitted tasks, which affects access to future work. Labor’s productivity is thus surveilled and disciplined exclusively through technology, both alienating workers and leaving them with no possible interaction with one another on the interface. Subtle forms of resistance, such as slowing the rate of work, are only detrimental to MTurk workers because they are paid by gig. Unlike Amazon warehouse workers, organizing through traditional strikes and unionization are for one complicated by the precarious nature of the work that classifies them outside of formal employment, while their web-based nature leaves them outside of the spheres of regulation altogether. If MTurk workers decide not to complete a task, it will simply be completed by someone else in the world on the platform – making this form of strike ineffective in disrupting MTurk. The global supply of labor where workers are interchangeable, combined with the precarious nature and the absence of a common workplace for these workers on the platform itself, are therefore among the crucial factors that undermine traditional organizing.

Though these workers cannot organize at the digital workplace in the same way that Amazon warehouse workers can, and are located in their own material realities with their own subjectivities, their organization takes an alternative form. Away from unions and industrial relations, workers instrumentalize the decentralized infrastructure of the Internet to rate and review requesters on Turkopticon, a Chrome extension created by Lilly Irani and Six Silberman, and use subreddits and forums such as Our Hit Stop to exchange tips and help others to find high-paying and credible requesters. Through these alternative fora, workers on digital labor platforms can advise one another and form solidarity through and on digital collectives.

While technology has always been instrumental in structuring working conditions, platforms demonstrate to us just how intrinsic to and inseparable from work technology-enhanced surveillance has become. At the core of this is algorithmic management. As I’ve described here, and cover in greater depth in my book through case studies on Amazon warehouses and MTurk, two dimensions—the nature of the platform and nature of the work—inform how platforms instrumentalize technology to differing degrees to mediate, manage, and surveil labor. While these cases studies underline how electronic surveillance diminish workers’ opportunities to engage in smaller acts of resistance, they also demonstrate how workers navigate this challenge and organize themselves in turn in both traditional and alternative ways. Grasping the power dynamics on the (digital) shop floor is a crucial step towards effectively regulating workplace surveillance and supporting workers in their struggles for better working conditions.