Taylor’s legacy: Sweat and exploitation in the digital factory
- review
- review of
Altenried, M. (2022) The digital factory. The human labor of automation. Chicago, IL/London: The University of Chicago Press.
Taylor’s legacy: Sweat and exploitation in the digital factory
Meet Flippy, the burger grilling robot working in a fast-food chain. Flippy inspects the meat patty with image and thermal sensors, decides when it is done, and turns it with a spatula. The spectacle of automation draws customers’ attention, yet the robot relies on human assistance. The rest is in the hands of employees: Burgers must be laid on the grill, wrapped, bagged, and passed to customers. Flippy vividly elucidates the shortcomings of automation through more or less smart machines and is one of many examples Moritz Altenried presents in The Digital Factory for how technology transforms labor in ‘digital capitalism’ [4]. Instead of replacing service workers with automation, robot work affords other, often unskilled tasks for human workers. Altenried’s depiction of contemporary capitalism distances itself from the idea of steadily increasing automation in the world of work. This includes going beyond analyses that predominantly emphasize the ‘creative, communicative, immaterial, or artistic features’ [5] of technologically changed labor relations. The author is not concerned with a critique of these approaches. He carefully notes that he is not ‘denying the importance of, for example, creative labor for the contemporary or ongoing processes of far-reaching automation’ [5]. The endeavor, rather, is ‘to provide a fuller picture’ of today’s forms of labor [160], including those which usually receive less attention in writings on digital capitalism. Accordingly, this book features workers sorting parcels in Amazon’s fulfilment centers, self-employed drivers transporting packages in their own cars, crowdworkers working in their living rooms and taking on micro tasks assigned via platforms for a few cents, and content moderators dealing with the psychological strain of keeping social media feeds clean. These ethnographically researched examples back the main argument that the logic of the factory still dominates capitalism. Digital technology, argues Altenried, ‘creates a set of very different labor situations, in which a new digital Taylorism exists alongside more autonomous forms of (immaterial) labor’ [17].
Analyzing the digital factory
Altenried situates his endeavor in the tradition of ‘Italian workerist Marxism’ [17] and quickly moves away from the industrial factory floor as a spatially enclosed place where workers assemble parts at a pace dictated by the assembly line. The author follows the decentralization of the factory, the effects of which he traces through three ‘vectors’ of analysis: ‘digital Taylorism’ [7], the ‘multiplication of labor’ [8], and ‘the reconfiguration of space through digital infrastructures’ [9]. The concept of digital Taylorism serves Altenried first of all to make quantifying work processes, their standardization and algorithmic monitoring describable. This is not about a ‘simple rebirth’ of Taylor’s scientific management, but about tracing ‘how digital technology allows for the rise of classical elements of Taylorism in often unexpected ways’ [7]. For example, there are employees in Amazon logistic centers whose job is to scan barcodes. While the barcode symbolizes the globally standardized registration of goods, the scanning process makes it possible to render work quantifiable by means of the items recorded. Based on this, management determines what is considered efficient work and creates specifications for how many barcodes must be scanned every hour. Such new techniques ‘to raise labor efficiency’ crystallize also in ‘the handheld scanner dictating workers’ pathways through Amazon’s distribution center’ [38]. The digital factory is characterized by the datafication of performance, the standardization of tasks, and continuous optimization, without the need for permanent monitoring by managers.
The second vector is the ‘multiplication of labor’ [8]. It figures as the sharpest distinction from the logic of the classic factory, for its digital counterpart does not produce an ‘industrial mass worker.’ [8] Digital technologies rather allow to access ‘a multiplicity of often deeply heterogeneous workers in multifarious ways.’ [8] The digital factory, for example the platform, allows to connect ‘a large number of workers who may come from different backgrounds, experiences, and locations’ [8]. Crowdwork on Amazon Mechanical Turk illustrates the point. Scattered around the globe, workers with Internet access log in and take on microtasks that are comparatively quick to learn: ‘categorize pictures, test software, transcribe audio recordings, or optimize search engine results.’ [93] Nevertheless, given ‘questions of nationality’ or ‘skills’, access to such microwork is unequally distributed, ‘and workers in the Global South earn on average less than their counterparts from the Global North.’ [118] Moreover, Altenried discerns a ‘literal multiplication of labor,’ meaning that many people work several jobs [8]. This refers not only to the fact that people can or must earn extra money with microtasks or work alongside their studies, but often do this in addition to performing care work. ‘Many women on crowdwork platforms’, the author summarizes his interviews, ‘explain that they had to leave their regular jobs to take care of chronically sick or elderly family members and resort to crowdwork to make up for lost income.’ [113] Regarding the multiplication of labor, digital technology allows for ‘the proliferation of flexible contractual forms such as short-term, subcontracted, freelance, and other forms of irregular employment.’ [9]
Besides digital Taylorism and the multiplication of labor, the third analytical vector is ‘the reconfiguration of space through digital infrastructures’ [9]. One manifestation of this is virtual migration. In the chapter on gaming, the author concentrates on so-called Chinese gold-farmers in the online game World of Warcraft. There are players in China who advance characters or collect digital goods and sell them for real money to others. ‘In most cases,’ Altenried observes, the gold farmers ‘will exploit one lucrative function or loophole in the game over and over.’ [68] Players toiling in organized groups in Chinese workshops ‘belong to the emerging digital working class of the Global South’. But within the landscape of the game itself, the gamers ‘are addressed as illegal migrants working in a space where others play’ – not work [74]. On the one hand, Chinese players represent ‘outsourced labor power in the periphery.’ On the other hand, the gamer is regarded as ‘a migrant doing the dirty work of farming while being subject to racist attacks within the digital sphere.’ [76] The term virtual migration aptly grasps the dual position of online workers: workers ‘migrate without migration.’ [76]
The not-so-hidden abode of ‘artificial artificial intelligence’
Altenried visits places where machines cannot substitute human decision-making. This applies to ‘gaming workers’ who are part of a game developer’s ‘quality assurance’ team and test versions of a game over and over ‘to search for errors’ and ‘report them to the software engineers’ [83]. Algorithms fail to reliably identify duplicates of products offered on Amazon and thus require human decision-making, and the same holds for Facebook’s content moderation. Given the ‘contextual nature’ of ‘cultural norms and practices,’ algorithmic decision-making flops [143]. This is particularly true when it comes to ‘hate speech or bullying,’ as the author explains: ‘the software has a hard time making sense of these highly contextual situations and is incorrect in the majority of these cases’ [143].
Such forms of human labor, according to a central theme of the book, are often made invisible. Content moderators are ‘crucial’, yet they are ‘mostly hidden from view of Facebook users and the public’ [140]. Also, crowdwork platforms provide ‘millions of hours of hidden labor that is necessary for algorithms that power self-driving cars or allow devices to understand human language’ [98]. At the same time, ‘the discourse around autonomous driving […] tends to overstate the ability of algorithms while obscuring the human labor” [100]. Besides ‘labor hiding behind the algorithmic search engine,’ there is ‘the labor of coding as well as maintaining the algorithmic architecture’ [127]. Gold-farming is depicted as a ‘shadow economy’ [166]. And the image of standardized shipping containers magically circling the globe without logistics workers must be countered by ‘a perspective that resists the picture of a seamless logistical harmony of endless circulation’ [22-23]. Clearly the book is committed to disclose the spaces in which workers toil. With this, Altenried is in good company. The growing interest in unveiling hidden labor relations transformed through digital technologies materializes in works on Uber (Rosenblat, 2019) and Deliveroo (Cant, 2019), on Amazon fulfilment centers (Beverungen, 2021; Hill, 2020) or Amazon Mechanical Turk (Irani, 2015; cf. Gray/Suri, 2019). It is situated within a discourse formation deconstructing myths of AI, shedding light on exploited workers and human decision-making (Moore/Woodcock, 2021; Smith, 2020; cf. Crawford, 2021).
But how invisible are these spaces of work? In the case of logistics and last mile delivery labour, Altenried notes that a glimpse at our cityscapes reveals to us an ‘army of delivery vans and couriers on bikes and scooters’ [49]. This ‘not-so-hidden abode of circulation’, as Annie McClanahan aptly puts it (2022, 315), is far from being invisible or obscured. It’s in plain sight. At the same time, Amazon fulfilment centers are placed ‘on the city’s outskirts’ [49], literally out of view. This proves to be a productive contradiction. What could follow from this is not to buy into the commonplace of hidden labor, but to refine what kinds of (in)visibility, (un)knowability, and (un)awareness of labor we can actually describe. Regarding the all-too-visible exploitation of delivery workers, the focus shifts towards consumers conveniently ordering goods online, who are probably aware of the exploitation of the delivery workers who will bring their purchase to their doorstep (cf. McClanahan, 2022; Neves/Steinberg, 2020).
Altenried’s book makes labor the subject of discussion and takes a critical look at places like sweatshops, distribution centers, and industrial ports. We know the tired faces of delivery couriers and curse at wrongly parked logistics vans. We may have heard about low-paid crowdwork or seen a documentary about the stressful filtering work of content moderators. And reports of Amazon employees being strictly disciplined now circulate in anecdotal form. The merit of this book is that it pulls together all the scattered notions of the effects of digital technologies on labor and makes their commonalities describable with digital Taylorism as a conceptual lens. The appeal of the work lies precisely in the combination of small-scale ethnographic work and looking at the whole in terms of how technologies are changing labor relations in global capitalism. The numerous examples reference interviews with workers, participant observation, ‘online (auto-)ethnography’ [12], the reading of patents and ‘reverse-engineer[ing]’ algorithms [14]. This combination of methods and well-arranged pieces make up the mosaic of the digital factory.
Connection points for further research
In some sections, more detailed explanations would have been appreciated. Even though the logistics chapter in particular talks about the global movement of goods as ‘anything but seamless’, [61] a stronger emphasis on seams and frictions would have rounded off the chapters further (cf. Gregson et al., 2017). This is somewhat due to the fact that algorithmic management is described as strict and total. Digital Taylorism characterized by ‘real-time control, feedback, and correction’ [159] sometimes appears to be too absolute. The gaps and breaks are missing, and the passages on (modes of) workers’ resistance are rather sparse – however, academic books might not be the most suitable place to share tactics of organization against labour regimes in the first place.
There are also conceptual imprecisions in talking about digital technologies, algorithms, infrastructures, software, and code. Here, reference to current positions from media studies might have sharpened the concepts. Though the book does cite code and software studies, these seem strangely outdated in light of recent in-depth research on datafication (cf. Thylstrup et al., 2021). Conversely, as The digital factory precisely analyzes how labour relations are reconfigured through technology, especially media studies could benefit from placing a stronger emphasis on labor regimes in descriptions of digital technologies.
Finally, but this is a matter of taste, a bolder positioning of the book would have been refreshing. Digital Taylorism, the argument goes, is only ‘one tendency among others’, not ‘the new hegemonic form of labor’ [17]. With similar caution, the author states that ‘this book provides a necessarily incomplete picture and perspective’ [16], humbly aiming ‘to provide a fuller picture’ of contemporary capitalism [160]. In contrast to works that have underscored the (sole) role of immaterial labor in digital capitalism, the book sees itself as ‘a supplement and moving of focus and perspective.’ [17] The chapter on logistics even states that it ‘provides an important appendix to many standard narratives.’ [25] Here, at least a conclusion that rejects techno-fetishistic analyses of digital capitalism which tend to neglect actual labor relations in favor of the marvels of automation would have been welcome. This is especially true of the much-cited surveillance capitalism. Grounded on behaviorist thinking, this form of capitalism merely focuses on a grim Google dystopia in which our data are scraped, sold, and used to alter our consumption preferences, but beyond that fails to pay attention to the very forms of labor Altenried describes in such minute detail (cf. Morozov, 2019). It’s capitalism without the digital factory. This also applies to theoretical efforts to grasp finance capitalism as the exclusive form of capitalism. Often presented as a radical break with and replacement of industrial capitalism, authors overemphasize the role of information and communication technologies and neglect forms of labor that are not executed in front of screens or done by algorithms (cf. Roberts/Joseph, 2015). In short: there are data flows and smart networks, but there is no exploitation and no sweat. Although Altenried’s The digital factory provides the solid ground for criticizing such approaches, to conduct this criticism is up to readers here.
In sum, Altenried delves into contemporary spaces of labor that are not exclusively limited to creative work or immaterial labor. Materials are transported across global supply chains, goods are processed in fulfilment centers, and food is delivered to the doorstep. Social media feeds are cleaned, duplicates on Amazon are found, and games are tested. All this requires human decisions and human hands. Accordingly, the first addressees of the book are those who merely focus on cognitive or creative work when it comes to technologically altered labor relations, or those who speak of the end of work with regard to automation. The digital factory is a recommended reading for all who want to gain a comprehensive insight into the changes in contemporary capitalism, especially because Altenried’s core arguments derive their persuasive power from the fact that the book remains so close to its case studies. The book avoids lengthy theoretical exegeses but is far from undercomplex, and places the conceptual framing of Marxist theory in digital capitalism primarily in the footnotes. The book’s readability with its varying depth and entry points makes it attractive to students of cultural studies, media studies, and social sciences. Altenried offers a well-written diagnosis of the present that vibrantly translates the logic of the factory into the present. Based on various spaces workers toil in, the book traces in detail the ruptures and continuities from the factory to the working conditions changed by digital technologies. Despite all the smart machines and promises of automation, the author convincingly emphasizes that human labor is by no means being replaced. ‘Today’s world,’ Altenried poignantly concludes, ‘is still a world of labor.’ [170]
acknowledgement
Thank you to Randi Heinrichs and Milan Stürmer for bringing the content of this review to life.
Beverungen, A. (2021) ‘Remote control: Algorithmic management of circulation at amazon’, in M. Burkhardt, K. Grashöfer and M. Shnayien (eds.) Explorations in digital cultures. Lüneburg: Meson press.
Cant, C. (2019) Riding for deliveroo. Resistance in the new economy. Cambridge, UK: Polity.
Crawford, K. (2021) Atlas of AI: Power, politics, and the planetary costs of artificial intelligence. New Haven: Yale University Press.
Gray, M. L., & Suri, S. (2019). Ghost work: How to stop Silicon Valley from building a new global underclass. Boston/New York: Houghton Mifflin Harcourt.
Gregson, N., Crang, M. and C. Antonopoulos (2017) ‘Holding together logistical worlds: Friction, seams and circulation in the emerging ‘global warehouse”’, Environment and Planning D: Society and Space, 35(3): 381-398.
Hill, D. (2020) ‘The injuries of platform logistics’, Media, Culture & Society, 42(4): 521-536.
Irani, L. (2015) ‘Difference and dependence among digital workers: The case of amazon mechanical turk’, South Atlantic Quarterly, 114(1): 225-234.
McClanahan, A. (2022) ‘Essential workers: Gigwork, logistics, and the sweated labor of circulation’, in M. Denecke, H. Kuhn and M. Stürmer (eds.) Liquidity, flows, circulation. The cultural logic of environmentalization. Zürich: Diaphanes.
Moore, P. and J. Woodcock (eds.) (2021): Augmented exploitation: Artificial intelligence, automation and work. London: Pluto Press.
Morozov, E. (2019) ‘Capitalism’s new clothes. Shoshana Zuboff’s new book on “surveillance capitalism” emphasizes the former at the expense of the latter’, The Baffler [https://thebaffler.com/latest/capitalisms-new-clothes-morozov].
Neves, J. and M Steinberg (2020) ‘Pandemic platforms: How convenience shapes the inequality of crisis’, in P. D. Keidl, L. Melamed, V. Hediger and A. Somaini (eds.) Pandemic media: Preliminary notes toward an inventory. Lüneburg: Meson.
Roberts, J.M. and J. Joseph (2015) ‘Beyond flows, fluids and networks: Social theory and the fetishism of the global informational economy’, New Political Economy 20(1): 1-20.
Rosenblat, A. (2019) Uberland. How algorithms are rewriting the rules of work. Oakland: University of California Press.
Smith, J.E. (2020) Smart machines and service work: Automation in an age of stagnation. London: Reaktion Books.
Thylstrup, N.B. et al. (eds.) (2021): Uncertain archives: Critical keywords for big data. Cambridge, MA: The MIT Press.
Mathias Denecke works at the Institute of Media Studies at the University of Bochum and recently published his dissertation Information flows in digital cultures (transcript, 2023). Research interests include infrastructures, media theory, and data labor.
E-mail: mathias.denecke AT rub.de