Ausgabe 1, Band 14 – März 2025
Arendt among the machines: Labour, work and action on digital platforms
Timothy Charlton
Minderoo Centre for Technology & Democracy
University of Cambridge
1. Introduction
As a microcosm of domination and resistance, work, that is the relationship between workers, their bosses, their products and the workplace, has been viewed as a mirror of society itself. For social theorists, the changing nature of work is a litmus test for political freedom in society. In The Human Condition (Arendt [1958] 2018), Arendt goes one step further by laying out the mechanisms connecting the political and the productive realms while strictly separating the two. In other words, why changes to the ways we labour and work affect our capacity for political action. And why a life of labour (i.e. domination) cannot simultaneously be a political one.
That the world of work is constantly (perceived to be) changing was as true in the 1950s as it is today. Alongside social and political developments, this is due to technological advances in the ways work is performed and organised. What stimulated Arendt’s imagination when she wrote The Human Condition was a world full of visible technologies, such as “giant computers”, early space flight, human-made satellites and the looming threat of automation (Arendt [1958] 2018, 172). Taken together, these achievements signalled the disruption of the established configuration of labour, work and action through machines. Technology would inevitably spur on the ongoing collapse of political action into necessary but trivial labour, diminishing the foundations for a political society. Somewhat tongue-in-cheek, she asks what should become of humans whose primary activity (their labour) had been rendered obsolete by the fruits of their own ingenuity in the society of labourers that has run out of labour1 (Arendt [1958] 2018, 5). She leaves no doubt as to her opinion that this development will not be positive. Without a thriving political life to fall back on and owing to humanity’s tendency to waste away its time with apolitical leisurely pursuits (Arendt [1958] 2018, 117–18)—labour’s less strenuous though equally apolitical cousin—the collapse of the Arbeitsgesellschaft would be a moment of extraordinary crisis. Given the rather gloomy introduction2, readers of The Human Condition might be surprised at the sprawling historical, social and political analysis of labour, work and action that follows it.
The threats of technology Arendt perceived were quite literally looming over her. They were massive in scale, omnipresent in narratives and imaginations and, in the case of nuclear fission, outright threatening and destructive.
Today’s defining technologies are radically different in this regard. What characterises algorithms, Artificial Intelligence (AI) and the cloud is their latent, invisible nature3. The technological threat has shifted from the outside world to within our pockets and homes onto the screens of smart devices (Runciman 2020). Digital platforms structure most aspects of our daily lives, from conversations with relatives and friends, through professional activities, to news consumption. According to the popular and academic opinion, we are also once again witnessing a transformative technological turning point—a new future of work—driven by digital platform technologies (e.g. McKinsey 2021). This so-called platform (or gig) economy promises immediate on-demand access to services and goods (e.g. Uber or Deliveroo). The dissolution of spatial and temporal constraints through platforms mean services can theoretically be performed remotely from anywhere (e.g. Upwork or Fiverr). Physical objects can be rented out with little effort (e.g. Airbnb) and social media platforms, such as TikTok or Twitter/X, create spaces for entirely new professions such as influencers or streamers. Whole enthusiast communities on topics ranging from gaming to citizens’ monitoring of nuclear proliferation4 have found their home on Discord, a chat platform offered as a service by a US company5.
Platforms are both the where and the how of activity. They fundamentally upend the timing, location and modalities of gainful activity. Early supporters of the gig economy spun a narrative of freedom and flexibility around it, heralding platform work as an alternative to grinding nine-to-five office work (Shibata 2020). As critiques of the platformised mode of production emerge and warnings are sounded about pervasive algorithmic surveillance (Bucher, Schou, and Waldkirch 2021) or the quasi-sovereign statelikeness of platform companies (Lehdonvirta 2022), i.e. their ability to dictate minimum wages and levy taxes, detailed examinations of their effect on political capacity are notably absent.
In this article, I develop an analytical framework of activity based on a reading of The Human Condition as an analysis of political freedom in a world characterised by technology. Arendt set out to understand how our ability to be political is anchored the ways we labour and work. I posit that for Arendt, the morphology of human activity is a heuristic device to understand the political capacity of society and, in the context of her biography, its proneness to totalitarianism.
What this article does not address is whether Arendt’s political theory, which draws heavily from Ancient Greek, especially Aristotelian, thought, is suitable for contemporary society. Neither does it attempt a critique of Arendt’s model or present a comprehensive collation of her unfolding thought throughout her opus. It deliberately sidesteps the complex and variable distinction between private, public and social spheres she develops across her writings. Instead, it highlights the often overlooked and arguably underdeveloped centrality of technology as a structuring force and source of power in Arendt’s thought.
2. The Activity Lens: Labour, Work and Action
In The Human Condition, Arendt builds her theoretical apparatus by juxtaposing concepts derived from etymological and anthropological sources. Necessity and freedom, private and public6, perishability and permanence or worldliness are the dichotomies through which she develops her argument. Societies slide between these indicators, in the sense that a totalitarian society is defined by the total absence of public life and political freedom whereas a utopian civic-republican democracy exhibits genuine participation in a public sphere. In totalitarian societies, words and deeds are not carried onwards and are thus denied the chance of attaining permanence. The spark of natality, where it appears, fizzles out inconsequentially. The governance of society is then determined not politically but a priori by the dominating ideology. Where society falls on these gradients is revealed, according to Arendt, through the configuration of activity among its population.
In her analysis, Arendt differentiates between three core activities that exclusively make up active life (Arendt [1987] 2018): labour, work and action. In contrast to contemplation, which is solitary, activity is always relational, i.e. it defines the way we are vis-à-vis other people and the world. Arendt constructs an anthropological grounding (Bowring 2011) for the three types of activity, labour, work and action, and illustrates them as ideal-typical animal laborans, homo faber and zoon politikon. These should be understood as stylistic devices and not as representative of certain groups of people. When Arendt speaks of the rise of animal laborans, she does not refer to the establishment of a labouring class as a sociological category (see also Levin 1979) but to the skewing of the tripartite configuration of activity towards labour. Put differently, the ideal mix of activity that prevails in society becomes tinged by labour.
Labour forms the basis of Arendt’s ontological hierarchy. In combination with its flip-side consumption, it represents the “metabolism of humans and nature” (Arendt [1958] 2018, 98, 121). Labour is the necessary part of the human condition in that, reflective of life itself, it enables us to survive. Its products are either cyclically lost in consumption (labour begets labour) or, in case they represent a surplus to bare necessity, drawn upon by homo faber or zoon politikon to fuel processes of creation or action (Arendt [1987] 2018, 296–98). This is illustrated by the individual, who, having laboured sufficiently, engages in political exchange, or the factory owner, who is freed from the necessity of having to labour themselves. As humans, we are always also “a labouring creature” (Arendt [1967] 2020, 180, my translation), so work and action, building on labour, can never be pure activities, but they can be understood in relation to labour as states of relative absence of necessity and perishability. Uncurtailed and left to its own devices, fundamentally private labour grows to fill out the space provided and subsume all activity in an endless production–consumption spiral. Political life thus depends on labour, but where labour is, politics is not.
Work, the domain of homo faber, is in contrast to cyclical labour a linear activity. Its means and end are clearly distinctive: a craftsperson works their material with the purpose of creating a use object and contributing it to the world (Arendt [1958] 2018, 136). The values of homo faber, stability and durability, are conferred onto the products of the work process, which in sum form the world of things, i.e. everything that structures lives from factories, through buildings and laws, to algorithms. This human artifice also provides the possibility for political speech and action to become virtuous and exert influence beyond the human lifespan by attaining permanence, i.e. by ossifying into the world (Arendt [1967] 2020, 95, 130). It is through this process that things attain a political role by proxy. This imbuing of values onto things (or the lack thereof) is a key analytical aspect of Arendt’s critique of capitalist modernity. In a market-driven mass society, where consumption preferences dictate the design of technology and the products it brings forth, the function of the shared world of things as a political space is compromised. If the imbuement process is polluted and public infrastructure is made up solely of perishable and undurable products, society is trapped in a vicious cycle of labour and consumption (Arendt [1958] 2018, 132).
The final and highest category of human activity according to Arendt7, action, is performed by humans qua zoon politikon8. It necessarily depends on plurality, i.e. being embedded in the “web of the acts and words of other[s]” (Arendt [1958] 2018, 179, 188), and on being in a conducive environment (the world). Action is simultaneously expressive, deliberative and narrative. In other words, through “word and deed” we reveal who rather than what we are (Arendt [1958] 2018, 176, 178–79). We are the tellers of our own story intent on effecting something among fellow humans in the world. To do this, we need to be free from necessity and simultaneously connected and separated from others in a way to create a space for the mutual examination of our acts. Using the table as an example (Arendt [1958] 2018, 51–52), Arendt conjures the image of individuals seated together as the archetypal infrastructure of the public sphere. Therefore, action essentially depends on labour and work. Surplus labour provides the necessary condition for freedom by momentarily absolving people from necessity, allowing them to work on the world and act in it.
This exaggerated depiction of the tripartite configuration of activity should not be misconstrued as simple categories into which various activities performed by individuals neatly slot. Most activity is complex and multifaceted, exhibiting blended characteristics of two or more modes of activity9. Instead, the tripartite model serves Arendt as an analytical device to identify how major historical developments affected the political capacity of society, e.g. by elevating structure (work) over action through radical planning10 or by collapsing work into labour, e.g. in Marxian “social praxis” (Benhabib 2020). By showing how the moving parts of activity and political freedom are connected, Arendt provides a model for future analyses and for interrogating how claims about radical change in one domain, e.g. a new future of work, might affect our political faculty and by extension, societies’ robustness against totalitarianism.
3. Technology & The Human Condition
Was Arendt a technological determinist?
The Human Condition is a book about humanity’s relationship with technology. Two major technological transitions play central roles in the argument advanced by Arendt. First, the “automation” of human activity kickstarted by the invention of the steam engine, the conclusion of which are the “mute robots” devised by engineers (humans in their capacity as homo faber) with the intention of reducing the burden of labour (Arendt [1958] 2018, 142). However, where machines take on “monitoring and control” over humans (Arendt [1958] 2018, 149), effectively dominating them, Arendt sees humanity itself demoted to a state of labour and consumption. The rhythmic alignment of humanity and machines is both a culmination of Marx’s metabolism and a reflection of the maximal, towering technologies that marked the 1950s. In the context of The Human Condition, this first view of technology as bringing forth dumb but strong robots, who are “like all machines, mere substitutes and artificial improvers of human labor power” (Arendt [1958] 2018, 172), is a polemic device and reads almost like a caricature. It serves as the antipode to the humanity of an idealised political society modelled on Graeco-Roman Antiquity.
Second, un-earthly (i.e. celestial or microscopical) technologies, such as satellites or biotechnologies allowing humans to “act into nature” (Arendt [1958] 2018, 269), signify an unshackling of human agency. Tinkering with life itself, the domain of necessity and labour, leads to a future where humanity can “channel the universal forces of the cosmos around us into the nature of the earth” (Arendt [1958] 2018, 150). While this has been described as a hybrid form of labour and action (e.g. Szerszynski 2003), I am more inclined to interpret it as a warning about the destructive unpredictability of human action amplified by modern technologies. The potential of humans to act into and alter their own basic activities also contains a redemptive spark—the possibility of changing things for the better against all odds.
The silhouetted and simplified descriptions of technology Arendt uses to build her case—that automation is a dangerous step towards an apolitical mass society caught up in perpetual labour and consumption—are what makes The Human Condition seem strikingly antiquated at times. Like Marx, she stereotypes technology to create a “dichotom[y between] society, social interests, and politics, on the one side, and technology, artefacts, and machines, on the other” (Pinch 1996, 34). However, she refuses to attribute agency to the artificial realm. Ultimately, it is always humanity acting into nature, just as it is humanity acting “against ourselves” (Arendt [1958] 2018, 323) through its latest technological achievements. In other words, technology is an extension of power (a tool for domination) recklessly wielded by unscrupulous individuals or movements for their personal gain. Though Arendt cannot be considered a strict technological determinist, she attributes a structural role to technology. Foreshadowing the idea of “autonomous technologies”, Arendt assumes the material properties of technology (and its structuring quality) ossify homo faber’s “values, beliefs, assumptions, privileges, and preferences, and once in place, are relatively closed to continued reframing or reinterpretation” (Lievrouw 2014, 8). Power relations are designed into the artifice of the workplace through technology, meaning the factory is by design conducive to labour, not to work, and the studio or workshop towards work, not labour. Studying how the modalities of work change is therefore particularly salient. It raises the question of whether platform technologies create environments conducive to labour or to work.
For Arendt, animal laborans and their factory are fully devoid of political faculty. True change in the form of political action can therefore never originate from this setting—it must be carried in from the outside. Collectively withdrawing the labour necessary to sustain a system, such as a profit-making enterprise, is only a political decision insofar as it precisely involves not labouring. Arendt’s understanding of labour is thus strictly functionally bounded, which explains the prevailing concern throughout The Human Condition that labour is polluting the other domains of active life.
In summary, technology plays an ambiguous role in The Human Condition in that it is a book about technology but not overly concerned with technology11. Technology is simultaneously central to the argument that shifts in the modalities of labour and work (through technology) diminish humanities political capacity and overstylised in the form of a polemical juxtaposition of the human and the artificial. The Human Condition provides a rudimentary analytical toolkit to interrogate technological developments in terms of their impact on our political freedom. In the next section, I apply this framework in the context of two case studies about digital platform labour.
4. Case Study: Unpacking the Gig Economy
The emergence of the gig economy is a salient form of platformisation in that the digital platform plays a particularly pronounced role in intermediating the productive relationship between clients and workers. In other words, it gives us the opportunity to directly study the technological changes to labour and work brought about by digital platforms. In this section, I present two cases of the gig economy. First, I explore the composition of activity on Upwork, an online freelancing or “crowdwork” platform and second, I present an analysis of the human labour underpinning cutting-edge AI technology through click work platforms. I draw on published secondary literature as well as my own previous research into online support communities for crowdworkers to explore how labour, work and action are performed on digital platforms.
Context: What is the gig economy?
The gig economy, where digital platforms facilitate and intermediate transactions between clients and services, emerged during the early 2000s from a pressure pot of sharing economy ideals, advances in internet platform technologies (under the banner of Web 2.0) and a scaling mindset inspired by venture capital (Fitzmaurice et al. 2020; Ravenelle 2017; Howe 2006). Despite reliable statistics about its uptake being scarce (not least due to their being jealously guarded by platform providers), gig work has become a labour market fixture, with millions presumed to be engaging in it at least occasionally (Schmidt 2017; World Bank 2019; Berg 2018). The opaqueness around the gig economy’s true size and scale has not stopped gig work from being proclaimed a future of work (Berg 2018; Vallas and Schor 2020). Encompassing local (e.g. taxi services) and remote types (e.g. online freelancing or click work), gig platforms are nowadays a global phenomenon. Remote gig work might include complex freelancing tasks as well as the notorious and emblematic cog-in-the-machine Mechanical Turk, a platform for exchanging menial activities for low remuneration (Howcroft and Bergvall-Kåreborn 2019). What defines remote gig platforms in general is their promise of compressing geographical distance, i.e. matching clients and anonymous labourers regardless of country or time zone.
Gig platforms frequently present themselves as replacing the organisation with the marketplace12 as the locus of productive activity. This vision is especially powerful in that it promises to funnel cheap labour power from the Global South towards clients from the Global North and hard currency the other way. Anyone with access to a laptop and an internet connection could self-select into global labour markets. This is presented as a form of global neoliberal development aid that incidentally also yields a profit (Baldwin and Forslid 2020; Kuek et al. 2015; The Economist 2019). Despite their pure market rhetoric, gig platforms are effectively intermediating the worker-employer relationship (Munoz, Dunn, and Sawyer 2022; Wood et al. 2019) using algorithms to allocate, monitor and control the work process and related issues, such as payments (Lata, Burdon, and Reddel 2022; Kellogg, Valentine, and Christin 2020). On gig platforms, workers are not afforded a clear-cut identity and often presented as freelancers, micro-providers or even entrepreneurs, deflecting employer-responsibilities (Degryse 2016) from the algorithmic matchmakers, despite their panoptic control of the work process: “If you use an app to go to work, should society consider you a consumer, an entrepreneur, or a worker?” (Rosenblat 2018, 8). Regardless of the answer, what becomes clear is that the gig economy is a far cry from a “society of jobholders” (Arendt [1958] 2018, 31) in that it shows none of the immediate traits of a job. The familiar fixtures of the workplace, such as managers, cubicles and front desks, have been melted into an amorphous digital work environment. Just so, I argue over the next paragraphs, have the power relations residing in these fixtures been concentrated in the hands of platform providers.
In the following sections, I apply an activity lens to two cases of gig work. Changes to the modalities of work are the analytical point of departure for Arendt’s critique of modernity in The Human Condition. Assuming gig work to be a future of work among others (Vallas and Schor 2020) and emulating Arendt’s approach, I ask: How are labour and work performed on digital platforms? and How does the platformisation of work and labour affect our capacity for political action?
First case: Algorithmic governance on Upwork
Gig work is presented as flexible for both workers and clients (Liang, Aroles, and Brandl 2022). Gig platforms provide the arena to optimally match clients’ needs to available workers irrespective of geography, solely according to price sensitivity, skills and quality demands (Munoz, Dunn, and Sawyer 2022; Wang et al. 2020). In principle, this is conducive to self-determined work (rather than labour) and gig platforms are indeed keen on portraying themselves as enabling technology, allowing workers to monetise their skills free from the constraints of a nine-to-five job. In this narrative, homo faber could transcend the limiting factors of their geography, such as corruption, lack of social services or low wages, and tap into the global market directly. In contrast, the lived experience of remote gig work is often defined by increased vulnerability and coercive algorithmic control (Anwar and Graham 2020; Wood et al. 2019).
Upwork, the merger of two pioneering gig work companies, Elance and oDesk, in 2013, sees itself as the culmination of a mission to “reduce […] friction in [the] job market”13. In the evolution from online job boards, through social media dedicated to professional life, to near-real-time curation of freelancers’ services and their allocation to clients, Upwork portrays itself as the final stage, an emerging fixture “that will be around in 100 years”. In its own view, it is the natural environment for natively digital “knowledge work”. The suite of services offered by Upwork goes far beyond providing a simple hands-off marketplace. By allowing freelancers to curate their profiles, managing a reputation mechanism of reviews and rating, algorithmically allocating tasks, verifying identities, dispensing payments and monitoring compliance through intrusive screen-grab software solutions, Upwork ties workers closely into a work environment where all basic relationships, such as reputation, trust and remuneration are intermediated by the platform. Despite its centrality to everything in the platformised work relationship, the platform also steps into an arbitrator’s role by resolving conflicts between clients and workers, a function that would conventionally fall to an external party. Since workers and clients may well be from entirely different countries, their membership of the platform can be the only mutual source of formal rules and regulations around the performed work. This accumulation of powers that were previously the prerogative of states, such as the de facto setting of a “global minimum wage”, has been recognised in the literature as platforms’ increasing statelikeness (Lehdonvirta 2022)14.
Since gig platforms often do not offer space for workers to interact freely without the fear of repercussions, gig workers frequently gather on external social media platforms, such as Facebook or Reddit, to discuss their work and any problems arising from it (Gray et al. 2016). Analysis of gig workers’ discourse on support communities dedicated to Upwork on the social media platform Reddit between 2012 and 2020 surfaced a range of issues15. First, it showed that despite a global rhetoric, workers perceived themselves as anchored in the US locality of the platform intermediary in terms of legal and financial obligations. Second, workers required a significant amount of labour-for-work (cf. Wood et al. 2019) to attract good gigs by gaming the allocation algorithm, e.g. by optimising their behaviour and profiles through trial-and-error, learning from others or by providing free or underpaid work to gain positive reviews and bolster their standing with the scoring algorithm. Third, the shared identity of workers in these communities revolved around their mutual platform membership rather than the nature of their craft. Finally, workers perceived as problematic the opaque governance of the productive process by platforms, including the setting of prices for certain tasks and even the removal of individuals from the market.
What emerges is an image of Upwork as a digital platform work environment, resembling a global factory rather than an intimate workshop. This new environment is superficially global but practically anchors workers in certain geographies. It provides no opportunity for meaningful exchange and learning among workers, who turn to external spaces to accommodate these needs due to fears of repercussions. The centralisation of power over reputation and payments bind workers to the platform, creating dependency and fostering necessity. In combination, these attributes suggest, from an Arendtian perspective, gig work is a misnomer. Gig labour would more accurately reflect its lived experience.
Second case: Invisible labour in digital supply chains
Gig platforms focused on AI, such as Appen or Samasource, provide a hidden-from-view interface between human workers and many other seemingly artificial or autonomous digital technologies, such as generative artificial intelligence or self-driving cars. They supply the manual labour that generates the training data for algorithms (Tubaro, Casilli, and Coville 2020; Tubaro and Casilli 2019; Irani 2015). Many of the foundational AI models currently dominating the news are based on a staggering amount of human activity, such as image tagging and annotation, generally classed as click work16, which is performed through relatively opaque digital platforms (Munn 2024; GPAI 2023). This fits a wider trend under which the “magic sauce” of many cutting-edge and superficially autonomous technologies is, contrary to what we are often led to believe, menial human labour.
Click work is deliberately omitted from the narrative around high tech. Venture capital investments incentivise high-tech digital solutionism, meaning the less visible the people feeding into such systems through platforms are, the better. The prominent poster child of generative AI, OpenAI’s ChatGPT, is reported to have relied on labourers in the Global South to perform psychologically challenging filtering and content moderation tasks for low pay17. The internet and digital platforms have indeed reconfigured labour, but they have done so by redistributing it globally (Munn 2022). As the creators behind sophisticated AI models and other technologies tell a story of liberating automation, i.e. of an infrastructure that makes certain types of labour obsolete, the peripheral platform labour is swelling. As cogs-in-the machine, humans are being worked on in “heteromation” arrangements (Márton and Ekbia 2021; Ekbia and Nardi 2014)‚ themselves becoming the components of a mechanistic system—something Arendt explicitly cautioned against when she observed how, in factories, the labourer “adjust[s] the natural rhythm of his body to [the machines’] mechanical movement” (Arendt [1958] 2018, 147).
Platformisation as labourisation
The two presented cases demonstrate how the quality of work is undermined through digital platforms. In the first scenario, gig platforms promise a novel flexible way of working according to one’s skills and volition. Liberated from temporal and spatial constraints, the remote gig economy claimed to facilitate the exchange of self-actuated work in the Arendtian sense. However, closer studies of gig platforms reveal that they are conducive to labour, not to work. If the boasts of platform providers, such as Upwork, are correct, and gig platforms indeed represent the natural space for digital work and they are, as the evidence suggests (Kässi, Lehdonvirta, and Stephany 2021), on the rise, then this represents a skewing towards labour in the tripartite model of human activity.
The second case demonstrates how through “tech-washing” the new digital infrastructure, such as generative AI, that should structure our being-together in the world is not only built on a foundation of labour but fully infused with it. The values of labour permeate our shared space in the shape of biases and inconsistencies. Its stability depends on the exploitation of many who are conveniently out of sight. Contrary to Arendt’s assumption that automation would eliminate labour, digital platforms are propelling it both in terms of distribution (across the globe) and scope (into the heart of contemporary technologies). The new human artifice carries traces of labour within it. Content generated by AI, for example, is interspersed with human labour and thus with the values of animal laborans.
5. Conclusion
This analysis showed how the Arendtian configuration of labour, work and action as measures of the quality of human activity retains its value even in a contemporary context. The consequences of the identified developments, the expansion of labour in scope and its diffusion into high tech, can only be speculated on. As gig platforms algorithmically govern production and individuals labour as component of a mechanistic system, they might be subjected to loneliness and isolation, qualities that we seem to intuitively associate with a networked society (Mason 2019) and which Arendt understood to be the preconditions for totalitarianism (Arendt [1951] 2017, 623). Other consequences of melting labour into work through sprawling global digital supply chains might be an erosion of stability or “world-alienation” of the public realm (Arendt [1958] 2018, 254). In contrast to Arendt’s professed fear that automation would expose humanity to mindless pursuits of leisure, platformisation has instead managed the reverse, by expanding labour into all activity.
From this insight we can draw normative conclusions about how we, as society, should react to the gathering storm clouds of AI and other future technologies. First, we should strip claims of automation from their veneer of craft and lay bare the actual compositions of labour and work that underpin them. This important work is done by journalists and academics, whose curiosity compels them to action, and who’s investigative work in this domain should be safeguarded and supported. If a technology claims to be foundational, i.e. infrastructural, then it must be our priority to ensure it is imbued with the appropriate values of stability and permanence that a functioning public sphere requires (Arendt [1958] 2018, 151–52). Second, if automation is not, in fact, a reduction of labour but its dispersal and amplification, then it is imperative to empower those toiling in the virtual factories of the cloud (and its very physical manifestations, predominantly in the Global South) to improve their labouring conditions. Self-directedness, pride in one’s work and the ability to participate in and shape the work process, i.e. the ‘human touch’ and means-end thinking of homo faber, are desirable qualities even for those at the most remote fringes of contemporary digital supply chains.
Finally, observing gig workers’ interactions on public social media communication platforms, such as Reddit, illustrates the various layers of activity at play in the platformised world. They labour on a dedicated gig platform, discuss (and organise) through speech and action on social media platforms and simultaneously work by contributing publicly available written traces and guidance to the world, which are picked up by generative AI18, algorithms and search engines, shaping our digital experiences. This study focused narrowly on the performance of labour and work on gig platforms, but how platformisation shapes other facets of activity, such as speech and action, is an area of active research (e.g. Charlton, Mayer, and Ohme 2024).
In this article, I attempted to transfer the tripartite concept of human activity developed by Arendt in The Human Condition to a contemporary setting. The cases have shown how the additional level of details provided by an Arendtian perspective are useful to analyse technological developments vastly different from those experienced by Arendt. This suggests they can form the basis for a more elaborate analytical and decision-making framework. Despite the outwardly pessimistic conclusions, the concepts laid out in Arendt’s The Human Condition lend themselves to an optimistic interpretation. Where society is structured to allow the spark of natality to fly, it can be redeemed. Digital life, including communities established and thriving on social media, various Open Source and Open Data movements and an increasing tendency towards unionisation among gig workers suggest that a grain of political freedom is always retained. Recognising the faults of the current techno-capitalist regime is a necessary step towards identifying and strengthening the genuine political institutions worthy of support. Using the activity lens to describe the composition of digital activities can aid this purpose.
6. Outlook: Towards an Arendtian Theory for the Digital Age
Finally, I addressed how a theory of labour inspired by Arendt can add a valuable perspective to the debate about the future of work in a society that perceives itself as undergoing a fundamental transformation. In this article I argued that the gig economy, one such frequently cited transformative development, is in fact an accelerated and covert form of labourisation. In doing so, I showed that forward-looking Arendt scholarship deserves a place at the societal table today. The activity lens I presented is only a first rudimentary step towards a more elaborate Arendtian framework for the digital age. It is based on my reading of Arendt’s central body of work in the 1950s as being fundamentally concerned with technology. Numerous avenues for further studies in this direction exist. First, in the face of the new imperious statelikeness of technology firms (Lehdonvirta 2022; Jin 2013), a rekindling with Arendt’s earlier study of totalitarianism—especially her investigation of the scramble for Africa and the role of equally statelike East India and Africa corporations—is in order. Initial engagements with this literature (e.g. Zuboff 2019) have reverberated strongly in the academic and mainstream literary communities, suggesting there is an appetite for further Arendtian approaches. Next, as generative AI tests the boundaries of truthfulness by producing almost indiscernible deep fakes that underpin disinformation campaigns on digital platforms, Arendt’s writings on truth and lying in politics (Arendt [1967] 2006; [1971] 1972) are worth revisiting. Open Source / Data approaches on the internet offer genuinely new possibilities of participation (Charlton et al. 2024), which resonate with the notion of council democracy. Finally, other future technologies, such as neural implants, beg the question of what happens when the domain of homo faber bleeds into the sanctity of contemplation, a field Arendt covers in her late work.
In summary, while the technological context of Arendt’s work on active life is somewhat dated, the ideas she presents are timelessly relevant. Given the richness of her multifaceted body of work, bringing Arendt back to the table by updating and reformulating her ideas for the digital age is undoubtedly worthwhile.
References
Anwar, Mohammad Amir, and Mark Graham. 2020. ‘Between a Rock and a Hard Place: Freedom, Flexibility, Precarity and Vulnerability in the Gig Economy in Africa’. Competition & Change, April, 1024529420914473. https://doi.org/10.1177/1024529420914473.
Arendt, Hannah. (1971) 1972. ‘Lying in Politics’. In Crises of the Republic, by Hannah Arendt, 1–47. A Harvest Book 219. New York, NY: Harcourt Brace Jovanovich.
———. (1967) 2006. ‘Truth and Politics’. In Between Past and Future, edited by Jerome Kohn, 223–59. New York: Penguin Books.
———. (1951) 2017. The Origins of Totalitarianism. London: Penguin Books Ltd.
———. (1987) 2018. ‘Labor, Work, Action’. In Thinking without a Banister: Essays in Understanding, 1953-1975, edited by Jerome Kohn, 291–321. New York (N.Y.): Schocken.
———. (1958) 2018. The Human Condition. Edited by Danielle S. Allen and Margaret Canovan. Second edition. Chicago: The University of Chicago Press.
———. (1967) 2020. Vita activa oder Vom tätigen Leben. Edited by Thomas Meyer and Hans-Jörg Sigwart. Erweiterte Neuausgabe. Serie Piper 31691. München: Piper.
Baldwin, Richard, and Rikard Forslid. 2020. ‘Globotics and Development: When Manufacturing Is Jobless and Services Are Tradable’. 26731. Working Paper. Cambridge, MA: National Bureau of Economic Research.
Benhabib, Seyla. 2020. ‘How to Read Hannah Arendt’s The Human Condition (1958) ?’ Critique (blog). 2020. https://blogs.law.columbia.edu/critique1313/seyla-benhabib-how-to-read-hannah-arendts-the-human-condition-1958/.
Berg, Janine. 2018. Digital Labour Platforms and the Future of Work towards Decent Work in the Online World. Geneva: International Labour Organization. https://search.ebscohost.com/login.aspx?direct=true&scope=site&db=nlebk&db=nlabk&AN=3109791.
Bowring, Finn. 2011. Hannah Arendt: A Critical Introduction. Modern European Thinkers. London, UK: Pluto Press.
Bucher, Eliane Léontine, Peter Kalum Schou, and Matthias Waldkirch. 2021. ‘Pacifying the Algorithm – Anticipatory Compliance in the Face of Algorithmic Management in the Gig Economy’. Organization 28 (1): 44–67. https://doi.org/10.1177/1350508420961531.
Charlton, Timothy, Anna-Theresa Mayer, and Jakob Ohme. 2024. ‘A Common Effort: How OSINT Communities and Journalists Interact on Digital Platforms’. International Journal of Press/Politics. https://doi.org/10.1177/19401612241271230
Dahrendorf, Ralf. 2007. Auf Der Suche Nach Einer Neuen Ordnung. Krupp-Vorlesungen Zu Politik Und Geschichte 3. Munich: C.H. Beck.
Degryse, Christophe. 2016. ‘Digitalisation of the Economy and Its Impact on Labour Markets’. 2016.02. Working Paper. Brussels, Belgium: ETUI. https://www.etui.org/publications/working-papers/digitalisation-of-the-economy-and-its-impact-on-labour-markets.
Ekbia, Hamid, and Bonnie Nardi. 2014. ‘Heteromation and Its (Dis)Contents: The Invisible Division of Labor between Humans and Machines’. First Monday, May. https://doi.org/10.5210/fm.v19i6.5331.
Fitzmaurice, Connor J, Isak Ladegaard, William Attwood-Charles, Mehmet Cansoy, Lindsey B Carfagna, Juliet B Schor, and Robert Wengronowitz. 2020. ‘Domesticating the Market: Moral Exchange and the Sharing Economy’. Socio-Economic Review 18 (1): 81–102. https://doi.org/10.1093/ser/mwy003.
GPAI. 2023. ‘Fairwork AI Ratings 2023: The Workers Behind AI at Sama. Report, December 2023’. Oxford, United Kingdom: Global Partnership on AI.
Gray, Mary L., Siddharth Suri, Syed Shoaib Ali, and Deepti Kulkarni. 2016. ‘The Crowd Is a Collaborative Network’. In Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing, 134–47. San Francisco California USA: ACM. https://doi.org/10.1145/2818048.2819942.
Howcroft, Debra, and Birgitta Bergvall-Kåreborn. 2019. ‘A Typology of Crowdwork Platforms’. Work, Employment and Society 33 (1): 21–38. https://doi.org/10.1177/0950017018760136.
Howe, Jeff. 2006. ‘The Rise of Crowdsourcing’. Wired, 2006. https://www.wired.com/2006/06/crowds/.
Irani, Lilly. 2015. ‘The Cultural Work of Microwork’. New Media & Society 17 (5): 720–39. https://doi.org/10.1177/1461444813511926.
Jin, Dal Yong. 2013. ‘The Construction of Platform Imperialism in the Globalization Era’. tripleC: Communication, Capitalism & Critique. Open Access Journal for a Global Sustainable Information Society 11 (1): 145–72. https://doi.org/10.31269/triplec.v11i1.458.
Kässi, Otto, Vili Lehdonvirta, and Fabian Stephany. 2021. ‘How Many Online Workers Are There in the World? A Data-Driven Assessment’, March. https://arxiv.org/abs/2103.12648v2.
Kellogg, Katherine C., Melissa A. Valentine, and Angéle Christin. 2020. ‘Algorithms at Work: The New Contested Terrain of Control’. Academy of Management Annals 14 (1): 366–410. https://doi.org/10.5465/annals.2018.0174.
König, Helmut. 1990. ‘Die Krise der Arbeitsgesellschaft und die Zukunft der Arbeit: Zur Kritik einer aktuellen Debatte’. In Sozialphilosophie der industriellen Arbeit, edited by Helmut König, Bodo von Greiff, and Helmut Schauer, 322–45. Wiesbaden: VS Verlag für Sozialwissenschaften. https://doi.org/10.1007/978-3-663-01683-0_17.
Kuek, Siou Chew, Cecilia Paradi-Guilford, Toks Fayomi, Saori Imaizumi, Panos Ipeirotis, Patricia Pina, and Manpreet Singh. 2015. ‘The Global Opportunity in Online Outsourcing’. World Bank, Washington, DC. http://hdl.handle.net/10986/22284.
Lata, L.N., J. Burdon, and T. Reddel. 2022. ‘New Tech, Old Exploitation: Gig Economy, Algorithmic Control and Migrant Labour’. Sociology Compass. https://doi.org/10.1111/soc4.13028.
Lehdonvirta, Vili. 2022. Cloud Empires: How Digital Platforms Are Overtaking the State and How We Can Regain Control. Cambridge, Massachusetts: The MIT Press.
Levin, Martin. 1979. ‘On Animal Laborans and Homo Politicus in Hannah Arendt: A Note’. Political Theory 7 (4): 521–31.
Liang, Yin, Jeremy Aroles, and Bernd Brandl. 2022. ‘Charting Platform Capitalism: Definitions, Concepts and Ideologies’. New Technology, Work and Employment 37 (2): 308–27. https://doi.org/10.1111/ntwe.12234.
Lievrouw, Leah A. 2014. ‘Materiality and Media in Communication and Technology Studies: An Unfinished Project’. In Media Technologies, edited by Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot, 21–52. The MIT Press. https://doi.org/10.7551/mitpress/9042.003.0005.
Márton, Attila, and Hamid Ekbia. 2021. ‘Platforms and the New Division of Labor Between Humans and Machines’. In New Ways of Working: Organizations and Organizing in the Digital Age. Technology, Work and Globalization. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-61687-8.
Mason, Paul. 2019. ‘Reading Arendt Is Not Enough’. The New York Review, 2 May 2019. https://www.nybooks.com/online/2019/05/02/reading-arendt-is-not-enough/.
McKinsey. 2021. ‘The Future of Work after COVID‐19’. McKinsey Global Institute. https://www.mckinsey.de/~/media/mckinsey/locations/europe%20and%20middle%20east/deutschland/news/presse/2021/2021-02-18%20-%20mgi%20future%20of%20work%20after%20covid-19/mgi_the%20future%20of%20work%20after%20covid-19_report_feb%202021.pdf.
Munn, Luke. 2022. Automation Is a Myth. Stanford, California: Stanford University Press.
———. 2024. ‘Digital Labor, Platforms, and AI’. In Introduction to Digital Humanism: A Textbook, edited by Hannes Werthner, Carlo Ghezzi, Jeff Kramer, Julian Nida-Rümelin, Bashar Nuseibeh, Erich Prem, and Allison Stanger, 557–69. Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-45304-5_35.
Munoz, Isabel, Michael Dunn, and Steve Sawyer. 2022. ‘Flexibility, Occupation and Gender: Insights from a Panel Study of Online Freelancers’. In Information for a Better World: Shaping the Global Future, edited by Malte Smits, 311–18. Lecture Notes in Computer Science. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-96957-8_27.
Pinch, Trevor. 1996. ‘The Social Construction of Technology: A Review’. In Technological Change: Methods and Themes in the History of Technology, edited by Robert Fox, 1:17–37. History of Technology 2. Oxford: Routledge.
Ravenelle, Alexandrea J. 2017. ‘Sharing Economy Workers: Selling, Not Sharing’. Cambridge Journal of Regions, Economy and Society 10 (2): 281–95. https://doi.org/10.1093/cjres/rsw043.
Rone, Julia. 2023. ‘The Shape of the Cloud: Contesting Date Centre Construction in North Holland’. New Media & Society, January, 146144482211459. https://doi.org/10.1177/14614448221145928.
Rosenblat, Alex. 2018. Uberland: How Algorithms Are Rewriting the Rules of Work. University of California Press.
Runciman, David, dir. 2020. ‘Arendt on Action’. Talking Politics | History of Ideas. https://www.talkingpoliticspodcast.com/history-of-ideas-1.
Schmidt, Florian A. 2017. ‘Digital Labour Markets in the Platform Economy’. Friedrich-Ebert-Stiftung.
Schwarz, Elke. 2014. ‘@hannah_arendt: An Arendtian Critique of Online Social Networks’. Millennium: Journal of International Studies 43 (1): 165–86. https://doi.org/10.1177/0305829814541505.
Scott, James C. (1998) 2020. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. New Haven: Yale University Press.
Sennett, Richard. (1977) 2017. The Fall of Public Man. New York: W. W. Norton & Company.
Shibata, Saori. 2020. ‘Gig Work and the Discourse of Autonomy: Fictitious Freedom in Japan’s Digital Economy’. New Political Economy 25 (4): 535–51. https://doi.org/10.1080/13563467.2019.1613351.
Szerszynski, Bronislaw. 2003. ‘Technology, Performance and Life Itself: Hannah Arendt and the Fate of Nature’. The Sociological Review 51 (2): 203–18. https://doi.org/10.1111/j.1467-954X.2004.00459.x.
The Economist, dir. 2019. The Global Gig Economy Is Helping People Lift Themselves out of Poverty. https://web.archive.org/web/20221201125140/https://twitter.com/TheEconomist/status/1090746719988523008/video/1.
Tsao, Roy T. 2002. ‘Arendt Against Athens: Rereading the Human Condition’. Political Theory 30 (1): 97–123. https://doi.org/10.1177/0090591702030001005.
Tubaro, Paola, and Antonio A. Casilli. 2019. ‘Micro-Work, Artificial Intelligence and the Automotive Industry’. Journal of Industrial and Business Economics 46 (3): 333–45. https://doi.org/10.1007/s40812-019-00121-1.
Tubaro, Paola, Antonio A Casilli, and Marion Coville. 2020. ‘The Trainer, the Verifier, the Imitator: Three Ways in Which Human Platform Workers Support Artificial Intelligence’. Big Data & Society 7 (1): 2053951720919776. https://doi.org/10.1177/2053951720919776.
Vallas, Steven, and Juliet B. Schor. 2020. ‘What Do Platforms Do? Understanding the Gig Economy’. Annual Review of Sociology 46 (1): 273–94. https://doi.org/10.1146/annurev-soc-121919-054857.
Wang, Blair, Daniel Schlagwein, Dubravka Cecez-Kecmanovic, and Michael Cahalane. 2020. ‘Beyond the Factory Paradigm: Digital Nomadism and the Digital Future(s) of Knowledge Work Post-COVID-19’. Journal of the Association for Information Systems 21 (November): 1379–1401. https://doi.org/10.17705/1jais.00641.
Wood, Alex J, Mark Graham, Vili Lehdonvirta, and Isis Hjorth. 2019. ‘Networked but Commodified: The (Dis)Embeddedness of Digital Labour in the Gig Economy’. Sociology 53 (5): 931–50. https://doi.org/10.1177/0038038519828906.
World Bank. 2019. World Development Report 2019: The Changing Nature of Work. Washington, DC: World Bank. https://doi.org/10.1596/978-1-4648-1328-3.
Zuboff, Shoshana. 2019. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. First edition. New York: PublicAffairs.
1It is worth pointing out that the German term “Arbeitsgesellschaft” evokes a different image than the English “society of labourers”, as is often the case with Arendt (Tsao 2002). The German term has seen a remarkable life of its own in the debate about the role of labour in society (e.g. Dahrendorf 2007, 62–63; König 1990).
2Paul Mason’s characterisation of Arendt as a “theorist of ‘what’s gone wrong and how humans [should] live […]” is fitting here (Mason 2019).
3This applies to most of society. There is a growing recognition of the physical and environmental footprint of the giant energy and resource guzzling data centres underpinning the technologies presented as shapeless clouds (e.g. Rone 2023).
4Arms Control Wonk or ACW and its associated Discord server is an example of a digital-native community gravitating around a mutual interest. See https://www.armscontrolwonk.com (accessed on 26.04.2024).
5Many prominent examples of digital platforms are incorporated in the US. However, of the named examples, Deliveroo is headquartered in the UK, TikTok in China and Fiverr in Israel, meaning this is not a universal rule.
6The quasi-private but outwardly public category of the social is not addressed here. It is best understood against the theatrical-performative backdrop of modern polite society beset by novel economic imperatives, where runaway internalised norms and codes might stifle meaningful interaction and structure a potential plurality into narratively homogeneous groups (see also Sennett ([1977] 2017) for an elaborate study of this phenomenon). In a fictitious labour-less labour society, the values and norms of labour may have been internalised to the extent that even in spaces where political action is in principle possible, its spark cannot fly. The social category has been previously applied to social media platforms (Schwarz 2014). Since digital platform are best understood as both the space and modality of activity, this approach is by itself not sufficient to determine their role in society. It must be complemented by a study of activity itself under platform conditions, which I attempt in this article.
7The normative assumption made by Arendt, inspired perhaps by Machiavelli’s virtù, is that good is what is political. Hence, action is the highest of the categories of activity, not in the sense of necessity but of value.
8Arendt’s extensive use of the word qua emphasises that animal laborans, homo faber and zoon politikon do not designate groups of people but states of being. For example, humans in the capacity of homo faber work, humans qua zoon politikon act, etc.
9For example, while painting a wall for money might be considered labour, painting it with a mural constitutes work by leaving behind an artefact that shapes our mutual world.
10The fallacy of ideology-driven societies is elaborated by Arendt herself in The Origins of Totalitarianism ([1951] 2017) and, in a more recent but illustrative example of runaway central planning, by James C. Scott in Seeing Like a State ([1998] 2020).
11In contrast, for example, to Heidegger in Die Frage nach der Technik, Arendt does not seem to assume that deep examination of technology itself is a worthwhile endeavour. Technology is only relevant in its capacity as an instrument of power, i.e. insofar it forms a nexus with activity and structures it.
12For example, Upwork say on their website: „We’re the world’s work marketplace” (www.upwork.com, accessed on 04.04.2024).
13Press release “Elance-oDesk Relaunches as Upwork, Debuts New Freelance Talent Platform” May 5, 2015. Available under https://www.upwork.com/press/releases/elance-odesk-relaunches-as-upwork-debuts-new-freelance-talent-platform (accessed on 25.04.2024).
14The accumulation of statelike powers by individual companies is not unprecedented. Earlier examples are, for example, company rule by the various East India and Africa companies.
15The results of this analysis were previously presented at the WORK and EGOS conferences in 2021 and 2022, respectively.
16During fieldwork in Nairobi, Kenya in 2019, I visited a so-called “click factory”. While the working conditions were not noticeably worse than other physical sites of platform work I had visited, what struck me was the obvious disconnect between labourers and the activities they were performing, which were broken down into repetitive tasks aimed at no recognisable end.
17This was reported by Billy Perrigo for TIME on January 18th, 2023. Available under https://time.com/6247678/openai-chatgpt-kenya-workers/ (accessed on April 24th, 2024).
18https://www.reuters.com/technology/reddit-ai-content-licensing-deal-with-google-sources-say-2024-02-22/ (accessed 31.10.2024)
