If early studies of information technologies in organizations were once a niche specialty in sociology, today’s ethnographers of labor cannot avoid the topics of digital systems, platforms, algorithms, or AI. One well-known challenge for the organizational sociologist studying such environments is the need to observe their continuities and discontinuities with prior organizational systems to best understand contemporary labor in digitally mediated spaces. Newcomers to these spaces must work studiously to avoid succumbing to the influence of hype around “the future of work” or comments from their respondents which naturalize these processes, such as by expressing the view that it is the Internet alone that has inherently changed the workplace, operating as a technologically deterministic force. Yet digitized workplaces also exacerbate another perennial challenge in organizational ethnography: the question of researcher positionality. Today’s workplaces are often strategically engineered in such a way as to maximize the distance both between employer and employee and between employees themselves. This poses a conundrum for the ethnographer, whose situated position and degrees of access inform what they can see and, accordingly, how they theorize.

Three recent books by emerging scholars of the future of work take on these challenges and address them with gusto: Hatim Rahman’s Inside the Invisible Cage: How Algorithms Control Workers [2024], Karen Levy’s Data Driven: Truckers, Technology, and the New Workplace Surveillance [2022], and Benjamin Shestakofsky’s Behind the Startup: How Venture Capital Shapes Work, Innovation, and Inequality [2024]. All three do an excellent job of cutting through the hype to interrogate the evolving textures of today’s digitally imbricated labor on the ground. As a result of their ethnographic positioning, these scholars ultimately develop contrasting approaches to algorithmic labor and the future of work. This presents implications for how we understand and study workplace dynamics today.

Hatim Rahman situates himself among distributed crowdworkers picking up jobs on a platform called TalentFinder. At first, people are attracted to the website because of its transparency in the matching process between workers and employers and because major companies use the service to staff temporary roles. Over time, however, the site transforms itself, enacting what Rahman calls an invisible cage: an algorithmically enforced “system of control that is objective, efficient, and scalable” [180]. Rahman’s focus on organizational cognition in this distributed and mediated environment enables him to identify the severe “asymmetry in the way organizations’ use of algorithms benefits actors” [181]. These algorithmic systems are not the only elements that are opaque to workers. Because TalentFinder’s platform, contracts, controls, and tools constantly change without warning or explanation, workers are subject to cognitive exhaustion as a result of the need for continual adaptation.

In an strategy that Rahman calls digital boilerplate creep, the company issues regular updates to worker contracts and terms of service as well as making changes to its website, ranking systems, and display functions. Acceptance of a prior contract presumes acceptance of all future terms. Workers are taken by surprise at the new policies and find the continual change disorienting. Since their reputations are bound up in the site’s ranking scores, they are forced to regularly shift their styles of interaction and strategies in order to ensure their success in the workplace. As in many ethnographies of digital crowdwork, we see that TalentFinder workers interact through online forums to exchange information and to speculate about what the algorithms are doing and why the company has implemented a particular change to the user interface or rankings. The workers—and Rahman, who also signed up to work on the platform—describe these changes as increasingly despotic and controlling, and aimed at ensuring that the employee’s voice within the firm is limited. In this way, Rahman’s work has much in common with prior studies of gig labor which have also found that platforms control and circumscribe that labor; for example, the work of Mary Gray and Siddarth Suri, Lindsay Cameron, Juliet Schor, and Steve Vallas.

Karen Levy’s fieldsite is also widely distributed: she studies truck drivers, whose industry is undergoing change through the enforcement of digital surveillance mechanisms. Levy focuses on the implementation of ELDs—electronic logging devices—in trucker cabs across the United States. In many ways, the ELD serves to take over some of the more tedious elements of the truckers’ work, such as logging hours and managing checkpoints. But truckers, as Levy explains in exuberant detail in Chapter 2, have a history and culture of rugged, frontier-like masculinity. As such, electronic monitoring devices challenge the trucker’s independence by introducing new forms of surveillance into the intimate space of the trucker cab, where the driver drives, eats, lives, and sleeps. Truckers do not take this intrusion lying down. Unlike Rahman’s, Levy’s interest is focused not on the caged and opaque elements of digitally managed labor, but rather on worker resistance.

Spending her time at truck stops and in trucker cabs across the country, Levy adopts an ethnographic lens that combines organizational and legal sociology with an eye to how policy intervenes in the corporate workplace. The ELD is “a legal, economic, and cultural object” and is understood differently by its various stakeholders [10]. Truckers interact around and with the ELD in dynamic ways that both destabilize and reinforce power relations. While Levy sees little distinction between government and corporations when it comes to the effects of surveillance upon workers [they are “pragmatically inseparable”: 76], she also dives into the many ways in which truckers, lawyers, and companies aim to enforce or resist surveillance regimes. Resistance, for Levy, is “a means of negotiating social power, rather than merely … bottom-up refusal” [14]. Truckers are not “docile bodies” [93] disciplined by the Foucauldian gaze of the tracking device or the company. As any trucker knows, a night in a freezer and a light tap of a hammer are all it takes to shut the ELD off for good! Resistance is part of the push and pull of employer/employee relations around voice and agency. This is not always negative for either the driver or the company. As Levy puts it, rephrasing an old aphorism in organizational sociology, “we depend on rule-breaking to make the world function” [4]. Levy’s interactional approach to digital tools in the workplace finds common ground with studies of professional contexts by scholars like Sarah Brayne, Angèle Christin, Kate Kellogg, and Melissa Valentine, in which digital systems are subject to local interpretation and contestation in the workplace.

While Levy clambered in and out of truck cabs and Rahman worked through online job listings, Shestakofsky was offered a position coordinating external workforces at an early-stage startup in San Francisco in the middle of an investment boom. The company, AllDone, produced a coordination platform much like TalentFinder, connecting local service providers with people needing odd jobs completed around their homes. From Shestakofky’s vantage point within the firm, the most important factor in how the work got done was not a platform, an algorithm, or a tracker: it was venture capital. Shestakofsky elaborates the organizational strain that investment places upon startups as forces of “lag” and “drag.” First, firms experience valuation lag: the “temporal and imaginative gap between a venture capital firms’ investment in a company and its ability to realize returns” [17]. As the company attempts to fill this lag, the pressure from investors to deliver ROI on a quarterly basis produces organizational problems, or drags. For instance, the distance between what the company’s product promises to do and the automation technologies available to it (technical drag) inspires the company to hire an outsourced team in the Philippines to do the work of the algorithmic matching and checking system. This frees the San Francisco team to appease their investors by scaling their local workforce. Thus software engineers in San Francisco spent the majority of their time not on software engineering but on interviewing software engineers. Meanwhile, the Philippines workforce must enact the system with human labor behind the scenes. This arrangement produces trust lag because the product does not work as intended, so the company outsources this problem to a call center team in Las Vegas, which operates as a customer help desk. Meanwhile, the internal team culture in San Francisco adopts the Silicon Valley ethos and its mottos “embrace chaos” and “ignore your customers,” so as to celebrate this organizational and technical arrangement as a form of success [156].

As Shestakofsky shows, AllDone relies on affective ties and momentum to ensure loyalty among these outsourced units. But as profits improve and the company approaches unicorn status, it becomes clear that the payoff will not be shared with those units. This produces organizational drag, as the workers upon whom AllDone relies to maintain the appearance of functionality realize they are not seen as valuable to the company after all, and never were. As the company celebrates its IPO, persisting global and gender inequalities across sites are amplified to a staggering degree to privilege those at the center of the corporation over those who have made the product work. Throughout Shestakofsky’s ethnography of the tech firm, the technology is sidelined even by those in the room, and is evidently seen as secondary to economic pressures: the overarching story is one of how capital shapes the company and its resulting technical capabilities. Thus, Shestakofsky’s economic sociology lens echoes the approach of those who focus on the political economy of technology, like Donald MacKenzie, Judy Wajcman, Marion Fourcade, or Jens Beckert, in that it observes economic activity through organizational and technical textures on the ground.

All three of these ethnographies successfully resist a bare deterministic narrative, instead focusing on how people get the work done. How each ethographer is positioned, however, influences their ability to theorize the technologies and textures of work they observe. For Rahman, as for his interlocutors, the algorithm is blackboxed and cannot be subjected to scrutiny:Footnote ¹ it is destined to remain “speculative” [Rahman 2024, 16] because the workers among whom he is situated are far removed from the company and its decision-making. This distance matters to theorizing: the “invisible cage” Rahman proposes is perhaps less a feature of a technology than of a working arrangement in which workers are distanced from the platform that coordinates their labor. For Levy, distance has always been a feature of trucking, and drivers have historically negotiated that distance culturally, politically, and technologically as a site of agency, resistance, and independence. As she explains in Chapter 4, the ELD is contested not merely with regard to its technical capabilities but because it has entered the field at a time when the trucking industry is consolidating. Levy’s truckers, longstanding independent contractors, are fighting oversight by government regulators and logistics corporations. The ELD itself, then, is a site on which truckers are fiercely defending their threatened economic independence. Shestakofsky’s privileged position at the center of the firm calls out the elephant in the room. He shows how the very workers that ethnographers are best able to study are so externalized that they are barely of any concern to the company at all. Instead, what we observe as technological and organizational effects—like surveillance devices or algorithmic opacity—are local instantiations of economic pressure upon the firm to consolidate capital. Shestakofsky’s case is the starkest in this regard, as to begin with, the technology isn’t even there. The platform possesses a Wizard of Oz setup, its matching “algorithm” composed of people in the Philippines imitating artificial intelligence to prop up an investment vehicle. As AllDone’s director of customer support and operations manager, Shestakofsky visits the Philippines and Nevada, and tries to bring up local workers’ concerns with the founders in San Francisco as investments ramp up. These appeals fall on deaf ears.

If the digital workplaces people find themselves in today—suffused with platforms, algorithms, and surveillance technologies—seem opaque and despotic, it is not because the technologies are so, but because they are the outcome of the pressures of platform capitalism.Footnote ² TalentFinder’s workers find themselves subject to boilerplate creep and changing modes of valuation not because the company is trying to control them, but because the company is restructuring to satisfy its investors. Like AllDone’s workforces in Nevada and the Philippines, the laborers on the TalentFinder platform are not the company’s chief concern, despite forming its core product. Its relationship with its investors is a far more pressing one, which configures technical and labor outcomes accordingly. The platform does not have to work; it needs only to work enough to scale up, produce ROI, and attain industry capture. These ethnographies also illuminate new facets of digital inequality. “Middle age women are what make AllDone work,” claims its CEO [131]; meanwhile, truckers’ devotion to a certain form of masculinity drives them to work overtime, without allowing trackers to stand in the way. Such systems capitalize on existing inequities, including global inequities among platform workers.

Where and how people work remains important. Yet unfortunately for the ethnographer, understanding the rationale behind these workforce arrangements is a difficult endeavor. We face an extreme case of convenience sampling when we position ourselves solely within a labor force whose work is mediated by distributed platform workers. Analytically, we are forced to take the perspective of those who are increasingly relegated to the outside; we must, like them, react constantly to changes rather than observe the levers of power in operation. Yet, increasingly, this perspective is the only one available to us for study, as companies, venture capital firms, and investors tighten their grip on the global economy, pushing labor ever further toward the margins. The mechanisms that drive these depleting and atomizing textures of contemporary work are at risk of becoming as opaque to the ethnographer on the shop floor as they are to the workers themselves.

The future of work, as Levy reminds us, is not some “dreamy or dystopian” vision of the future, nor is it “some distant or discrete mode of social organization so unlike the one we have today” [77]. Indeed, the future of work is not a technology at all, but an arrangement—of capital, bodies, and technologies—as these textured ethnographies illuminate. As we put our best efforts into understanding the digital economy from the ground up, we would do well not to naturalize the effects of platform capitalism as a unique form of labor, as “the future of work,” nor as intrinsic properties of a technology like an algorithm or AI. To do so would be to mistake the explanandum for the explanans; to reproduce a form of technological determinism by virtue of ethnographic positioning. If anything, these ethnographies reinforce the mutual entanglement of these two enduring challenges in studies of workplace technologies. We must both grapple with the strategic position of the organizational ethnographer and resist the ready narrative of technological determinism if we are to better understand the many futures of work.