Moritz Altenried’s The Digital Factory: The Human Labor of Automation (2022) accomplishes in just under two hundred pages what many other books twice that length have struggled with: assembling a concise yet readable introductory map to the global, fragmented, and too-often hidden landscape of digitally-mediated capitalism. In this evolving terrain of accumulation, heterogeneous groups of workers in data centers and fulfillment warehouses, private homes and mobile delivery vehicles, and yes, sometimes even factories, create the kind of value that fills in the cracks of culture and commerce in the ever-changing information internetwork. These largely invisible laborers move products across the “last mile” of global supply chains, find lurking bugs in the last days of crushing software development cycles, eliminate illegal and traumatizing imagery to sanitize social network feeds, and tag new knowledge destined to train the machine-learning systems of the future.

Altenried’s tour through this geography stops at all the main attractions one would expect: Alphabet’s “Googleplex” campus in Silicon Valley, where subcontractors wear different color badges than salaried workers; Amazon’s “fulfillment centers” across the globe, strategically located where the consumer homes are close but the rents and wages are low; UPS delivery trucks across the US routed centrally by a proprietary “On-Road Integrated Optimization and Navigation” (ORION) system; marquee gaming studios where “crunch time” before a major version release means long hours and sleeping in the office; and private homes where “people as a service” balance child care with online microwork doled out by social media platforms. These are all, as Altenried frames it, “sites that may not always look like factories, but where the logics and workings of past factories are very much present, often accelerated by the increasing pervasiveness of digital technology”.Footnote ¹

Conceptually, these are “all sites where digital technology produces labor relations” and also sites where “Technology is materialized social labor and is as such a product of the social relations that brought it into existence”.Footnote ² That dialectic between technologies, sites of labor, and labor relations repeats again and again: “similar contractual arrangements, a curious return of piece wages, labor management software operating according to similar parameters, similar questions of space and infrastructure in relation to digital economies” […].Footnote ³ And through this tour, Altenried has a keen eye for the contradictions of these digital labor relations: “Repetitive yet stressful, often boring yet emotionally demanding, requiring little formal qualification yet oftentimes a large degree of skill and knowledge, and inserted into algorithmic architectures not yet automatable (at least for now) – these segments of labor are a crucial part of the political economy of the present.”Footnote ⁴

The resulting book is extraordinarily well-written – clear, engaging, and carefully connected to some of the most interesting research across interdisciplinary fields of political economy, human geography, and science and technology studies. About a dozen interviewees scattered among the different work sites provide a solid empirical grounding (though not a level of variety or representativeness that would be found in a modest-sized case study of any of the sites alone). The detailed narratives and first-person accounts of the sites of work that Altenried investigated (often over a period of years) are supported but never overwhelmed by the theoretical framework he uses to identify patterns and frictions in the digital labor process. And Altenried brings in just enough local context and historical background to situate each site in a wider research conversation on labor, technology, and geography. I very much appreciate this combination of historical underpinning and heterogeneous case examples. Drawing mostly on secondary accounts, and gesturing both to the most recent theoretical literature and the classics of Marxian scholarship, this wide-ranging narrative makes the book an excellent choice for a graduate seminar or upper-level undergraduate course on labor in the new global digital economy.

For scholars of networked information labor, the book also offers an interesting puzzle in how to frame new technology in both contrast and continuity with the old. Altenried conceptualizes the overall lessons of his book using the metaphor of the late-nineteenth and early-twentieth century factory in a way that attempts to connect the historical traditions of the technological intensification of production under Fordism, the migration and massification of labor under urbanization, and the decomposition and control of work tasks by management under Taylorism. By updating these ideas for the digital age, Altenried illustrates how the continuing imperatives of capital accumulation still apply under new spatial, temporal, and technological relations of production. For example, Altenried writes of “digital Taylorism” as seen in platform companies which enforce “standardization, decomposition, deskilling, automated management and human computation, algorithmic cooperation, digital measurement, and surveillance of labor.” In other words, “digital technology in particular allows for the mobilization, renewal, and recombination of crucial Taylorist principles in novel ways and contexts”.Footnote ⁵ In this conceptualization, “Digital Taylorism is hence no longer bound to the disciplinary architecture of the factory,” so “today’s digital factory itself can take many different forms”.Footnote ⁶ Thus Altenried argues we must “theoriz[e] contemporary capitalism not by the factory’s end, but by its transformation, multiplication, and generalization”.Footnote ⁷

That metaphor of the digital factory, though, is also the main limitation of Altenried’s book. When Altenried argues that “digital capitalism is not characterized by the end of the factory, but by its explosion, multiplication, spatial reconfiguration, and technological mutation into the digital factory”, I’m not sure who he is arguing against – who is saying that digital capitalism heralds the end of the factory? – or how much work that metaphor ends up doing for the analysis. Altenried writes that we should understand the prototypical “factory” as “an apparatus and logic for the ordering of labor, machinery, and infrastructure across space and time”.Footnote ⁸ I agree; it seems to me that Altenried is really arguing for a conceptual redefinition of the modernist factory as a series of technological and social innovations and apparatus that are meant to assemble, enable, intensify, monitor, and control human labor in the reproduction of further technological innovations and apparatus (from steel parts and family automobiles to silicon microprocessors and personal computers). But one could use this same technique instead with the metaphor of the nineteenth- and twentieth-century office, itself a series of technological and social structures that are meant to assemble, enable, intensify, monitor, and control human labor in the production of further social innovations and apparatus – insurance policies, financial instruments, health care interventions, and entertainment content. The “digital office” of the present day could be a similarly generative metaphor for exploring digital capitalism – after all, Taylorism was just as enthusiastically applied to the work of clerks and stenographers, typing and filing, as it was to the moving assembly line.

Other critical internet and labor studies scholars have invoked the digital factory metaphor as well – particularly Trebor Scholz in the 2013 edited volume, Digital Labor: The Internet as Playground and Factory.Footnote ⁹ But that formulation was meant, in part, to highlight the contradiction between freely-chosen hobby versus rigidly disciplined work which characterizes many of the hidden but crucial labor roles that both unpaid users and microwork contractors often assume in the online platform environment – chat moderator, wiki editor, or even paid content creator. And the “digital factory” has a more literal meaning for global consulting firms like McKinsey & Company who refer to it as “Industry 4.0” or the “Fourth Industrial Revolution” to mean the “next phase in the digitization of the manufacturing sector, driven by disruptive trends including the rise of data and connectivity, analytics, human-machine interaction, and improvements in robotics”.Footnote ¹⁰ In many ways, what Altenried is actually doing with his narrative is uncovering the ways in which the digital factory itself is an incomplete concept, almost always requiring us to look for the external and contingent labor support hidden just outside of its supposedly totalizing network of logistics and robotics and algorithms.

Altenried acknowledges many of these limitations of the digital factory metaphor in other parts of the text. For example, he notes “a crucial difference to traditional Taylorism, namely, the digital factory produces no digital mass worker like the industrial mass worker”.Footnote ¹¹ This is where the factory metaphor and neo-Taylorist framing might limit analysis. While a persistent organizational and managerial ideal for heterogeneous, far-flung, round-the-clock, and contingent armies of “digital factory” workers might be complete surveillance and control at the smallest level of work output, just like in an “analog factory”, the tasks required of these workers nevertheless demand a certain agility, creativity, and freedom of action that eludes complete rationalization. This is precisely why assembling workers who represent a diversity of situated backgrounds, cultural experiences, and soft skills is so crucial: some degree of customer or client empathy, cultural expertise and interpretation, and emotional labor are often necessary for these jobs. Workers subject to the discipline of the “digital factory” must often simultaneously use their own judgment in dealing with clients across global cultural barriers when the commodity being exchanged is the kind of symbolic meaning that is circulated on social media systems, negotiated in help-line conversations, or embedded in machine learning data sets. For example, Altenried shows us that “[F]actors such as education, language, and culture […] have made the Philippines a global center for digital social media content moderation. Its colonial and postcolonial history has produced an English-speaking, Catholic workforce well versed in North American culture, and hence very much qualified to perform content moderation for Western social media”.Footnote ¹²

Neither “digital factory” nor “digital office” seem to capture the full ramifications of these sorts of technologically-mediated, spatially-fragmented, and institutionally-complicated social relations of labor, where the structures of time/space/wage discipline must also allow for the creativity of culture and empathy to emerge in the course of that labor. I fear that sometimes the modifier “digital” as applied to a historical concept is our interim strategy for developing new conceptualizations of our contemporary algorithmic society, much like the modifier “post” applied to modernity served in the 1980s and 1990s. I have not done much better in my own writing – choosing instead to focus on the concepts of “visibility” and “invisibility” to call out many of the exact same trends that Altenried covers in his guide to the same landscape. Perhaps one advantage of the “hidden labor” approach is that it reminds us how these workers are not only invisible to clients and consumers, let alone to investigators, scholars and policy-makers, but also, largely, to each other – except occasionally through the use of similar ubiquitous digital technologies for information exchange and self-organizing across time, space, and organizational boundaries. That condition of production seems to slip free of both the factory and the office metaphors, though substituting something like “networked capitalism” – another popular rhetorical strategy from the 1990s and 2000s – does not seem any more satisfying.

The provisional answer to some of these questions will probably come from the next generation of scholars to investigate this new map, and Altenried’s account itself points to several great examples of future work. For example, whether conceptualized as factory, office, or invisible network, each of those metaphors feels inadequate when trying to document and analyze the actual time and space trajectories of people’s up-and-down work lives in this new economy, from gig work to contract work to “permanent” positions and then back again. Or people's cycles of professional education, technological work, planned obsolescence, and remedial retraining or reschooling. Or even people's real and virtual patterns of physical and “virtual” migration in search of both remunerative work and effective education, either from one digital platform to another, or from one political entity to the next. As Altenried writes, “[V]arious factors, including flexibility, precarity, pay, and language determine how platform work is in a specific way incorporated into a stratified labor market, tied to diverse migration routes and projects, bridges waiting periods at other jobs, or complements other wage labor or employment forms”.Footnote ¹³

But all these questions just reinforce the richness of Altenried’s narrative. Even his short but provocative epilogue – connecting the analysis in the book to the global COVID-19 pandemic – demonstrates the conditions of crisis under which the invisible “last mile” digital factory workforce of algorithm-managed gig laborers suddenly becomes visible (and even more profitable). A next step might be connecting the more privileged digital office workers who moved to remote work based in their homes – despite the challenges of this transition, nevertheless finding temporary pandemic security at no loss of hours or salary, unlike the other digital factory workers – to the many still-in-place tech support and troubleshooting workers necessary to enable this mass exodus from the workplace. And the impending shift of countless people and jobs in and out through new cycles of artificial intelligence and machine-learning augmentation and replacement will complicate such life courses of employment and schooling even further. This set of relations and connections is important terrain for future scholarship and interdisciplinary exchange. No matter what metaphor for these patterns we might find productive at any given moment, Altenried’s wide-ranging story about the kind of human labor often necessary to (but simultaneously often fragmented, devalued, and hidden by) our new global digital infrastructures is a powerful opening map for our investigations.