The AI Fix?

The literature on artificial intelligence (AI) and algorithms is abundant in recent years, to say the least. Although turbocharged by the media and scholarly hype, this proliferation still speaks to very real processes of deployment of algorithmic technologies in numerous spheres of activity, with overwhelming impact on society. We began our research project on algorithms three-and-a-half years ago with a hypothesis: we cannot understand this accelerated diffusion of AI in society without understanding the reconfigurations of contemporary capitalism, and vice versa. We have since become convinced that these reconfigurations require the development of new concepts, rooted in research deploying a broad interdisciplinary and critical approach on multiple aspects of social reality: political economy, the state, subjectivities, nature, technology, international relations, labor, culture, time and temporality, ethics, social reproduction, power and resistance. We have summarized our findings over twenty theses, in a volume just published in French.¹ This article aims at a brief presentation, for anglophone readers, of our approach of what we understand as algorithmic capital, and an exploration of some of its impacts on the sphere of social reproduction. We hope to contribute to discussions and debates on the transformations of contemporary capitalism.

The first decade of the twenty-first century, witnessed the convergence of two distinct yet related phenomena. On the one hand, the emergence of cloud computing, big data, smart devices, social media, new business models in Silicon Valley, and machine learning and deep learning in the field of AI created a technoscientific and economic basis allowing for novel forms of capital accumulation via data extraction and the spread of algorithmic predictive machines. On the other hand, the Great Recession of 2007–09 severely disrupted an already exhausted model of neoliberal financial capitalism.²

While the collapse of global finance did not put an end to financialization, it ushered a certain shift in investment fluxes from financial products and speculation over to the tech sector, especially high tech and AI. Subsequent quantitative easing policies pursued by several central banks were also instrumental in redirecting potential high return investment opportunities towards high-tech endeavors. Tax evasion and cash hoarding by major tech companies also fueled riskier and bolder investments in new techs.³ Digital platforms, the high-tech sector, the so-called collaborative economy startups such as Uber, Airbnb, and other Silicon Valley inspired innovations garnered tremendous momentum (as well as venture capital) in the 2010s and took over as a spearhead of capitalist growth in the US and elsewhere.

We see the economic recovery of the 2010s largely as a technological fix to the crisis of the globalized neoliberal model. This fix was material–ideological. “Californian ideology” and the discourse of “technological solutionism” of Silicon Valley proposed to solve neoliberalism’s problems with tech: gig work for the unemployed, health apps instead of accessible healthcare, “connectivity” against isolation, smart monitoring of resources to renew growth and fight climate change, and so on.⁴ But while this fix helped put capital back in the saddle of profitability, it also helped establish a new logic of accumulation.

The tech sector, boosted by the monetization of major AI and algorithmic innovations, became a new engine of economic growth, gradually replacing the logic of financial accumulation with an algorithmic logic based on data extraction and behavior prediction through algorithmic machines. After 2008, Big Tech companies took over as some of the largest companies in the world, leapfrogging and outperforming big extractive, industrial and financial companies, before literally deploying their algorithmic logic and new model of accumulation back into these other firms and sectors. It is increasingly difficult to find economic sectors and markets not being reconfigured by big data and algorithms today.

We thus argue that the emergence of big data and the rapid deployment of AI via advances in machine learning and deep learning, coupled with the 2007–08 financial crisis, brought about a new stage of capitalism based on a specific logic of accumulation and new form of power, which we have dubbed algorithmic capital. This term refers to a multidimensional phenomenon: a formal logic, a dynamic of accumulation, a social relation and an original form of power based on algorithms. By contrast, the expression algorithmic capitalism refers to the social formation as a whole—the historical articulation of a mode of production and an institutional context, what Nancy Fraser calls an “institutionalized social order.”⁵ We can therefore distinguish between the capitalist society we live in, and the specific capital social relation that unfolds in a multitude of ways.

“Algorithmic capitalism” is neither the first nor the sole attempt to conceptualize the mutations of contemporary capitalism.⁶ The dazzling acceleration of technological development and the emergence of a globalized information economy in recent history have given rise to a wealth of terminology in the critical literature. Capitalism has been, among others: “informational,” “digital,” “communicational,” “cognitive,” “attentional,” “cybernetic,” “platform,” “surveillance,” each of these terms alternatively emphasizing different dimensions of these historical developments.⁷ This semantic proliferation is undoubtedly symptomatic of fundamental mutations in capitalism.

The point here is not to add a new label or buzzword to the pantheon of theories of capitalism just for the sake of it: we have become convinced throughout our research that the reconfiguration of the fundamental social relations that connect capital accumulation with algorithmic technologies requires us to adapt critical concepts accordingly. Indeed, we argue that the algorithm is the structuring principle of the new regime of capitalist accumulation that builds on, re-articulates and ultimately transcends neoliberal capitalism. Since we cannot address all the aspects of this transformation in this article, we focus below on the transformations in the sphere of social reproduction triggered by the introduction of AI technologies into the home, a process occurring most extensively in middle class and rich households in the Global North.⁸

Algorithms in the New Regime of Accumulation

While scientific literature has produced a constellation of appellations that emphasize different aspects of contemporary capitalism, the concept of “algorithmic capital” is more precise and offers key heuristic advantages for grasping the logic of the new regime of capitalist accumulation. The production and deployment of algorithmic technologies in the social field that extract data indiscriminately from labor time, leisure time, and reproduction time, disrupt the essential equation between labor time and production of exchange value that endured since the dawn of capitalism.⁹ As new modes of valorization attach to extraction, treatment, and transformation of data into different types of assets and commodities, the cogs of capitalist accumulation become entrenched in algorithms; be it in production, circulation, or consumption. Many new critical approaches in political economy tend to focus on the novelty of the data economy and the new productive relationship between “users” and capital. Encompassing these novelties, the concept of algorithmic capital, however, keeps labor and production at the heart of the analysis.¹⁰ It also identifies what, in our view, constitutes the core of the logic of accumulation in capitalism today, and of the social relations that unfold within it: the algorithm.

Building on a rich critical literature, we locate ways in which, from its central role in capital accumulation, algorithmic logic comes to structure and mediate key social processes.¹¹ First, the algorithm becomes the dominant mechanism for the allocation of digital labor. Second, the algorithm, or algorithmic accumulation, becomes the dominant mechanism in determining the production process. Third, algorithms and data assets become a central object of competition between capitalist enterprises. Fourth, algorithms mediate social relations, notably through social media and digital platforms. Fifth, algorithms mediate access to information and collective memory.

Sixth, algorithms generate revenue by participating in the production of goods and services, or as a mechanism for extracting differential rent. Seventh, beyond private companies, many organizations and social spheres, ranging from public authorities, police services, military agencies to NGOs, healthcare, education, transport, public infrastructures use algorithmic technologies to exercise their power and/or pursue their activities. Eight, algorithms reconfigure social temporalities. For all these reasons we see algorithms, far more than “the digital,” “cognition,” “surveillance,” or anything else, as the heart of the new regime of accumulation: they mediate social relations, increasingly preside over socio-economic (re)production, and disseminate their predictive logic throughout society.

We note several commonalities between our conception of algorithmic capitalism and Dyer-Witheford, Kjøsen, and Steinhoff’s theory of AI-Capitalism.¹² These authors highlight the decisive role of the general conditions of production in the development of capitalism, that is, the technologies, institutions, and practices that constitute the environment of capitalist production in a given place and time, gradually becoming essential factors common to all production. Railroads, telephones, and electricity are just a few examples of general conditions of production that have had a major impact on forms of production, distribution, and consumption over the history of capitalism. Our era, they argue, is characterized by the rise of AI, algorithms and platforms to the rank of such general conditions of production. Dyer-Witheford, Kjøsen, and Steinhoff proceed to a helpful periodization of capitalist development, but their divisions tend to remain mechanistic and techno-focused: in addition to technological changes modifying the general conditions of production, social and institutional dimensions must also be taken into account to explain the passage, always complex, multifaceted, and nonlinear, from one stage of capitalism to another. ¹³

Capitalism manages to stabilize itself and overcome its contradictions, for a time, through specific institutional configurations.

A useful approach to complement our theorization is found in regulation theory, which proposes that capitalism manages to stabilize itself and overcome its contradictions, for a time, through specific institutional configurations: wage relations, monetary and financial regimes, innovation systems, state regulations, and international relations.¹⁴ Taking these broader institutional and social elements into account enables us to think in terms of a stage of capitalism or regime of accumulation: a “set of regularities ensuring a general and relatively coherent progression of capital accumulation,” a “relatively stabilized institutional matrix” that enables the reproduction of capitalism in a particular configuration over a given period.¹⁵ The regime of accumulation is itself encompassed within a mode of regulation: a “set of procedures and individual and collective behaviors” that reproduces social relations, sustains the regime of accumulation, and makes individual decisions compatible with dynamic adjustments of the system.¹⁶

The combination of a regime of accumulation and a mode of regulation constitutes a mode of development, for example, Fordism. Using this conceptual framework, we argue that algorithmic capitalism represents a genuinely new regime of accumulation, based on the predominance of algorithmic capital and encompassed by a mode of regulation based on algorithmic governmentality.¹⁷ Algorithmic governmentality, also known as algorithmic governance or algorithmic regulation, plays a key role in regulating social relations and behavior.¹⁸ It steers the regime of accumulation, coordinating various decentralized decisions through algorithmic decision-making systems, nudges, predictive sanctioning systems, recommendations systems, data-driven performance management, preemptive regulation systems, and so on.¹⁹ We do not have the space here to develop this point more fully, but algorithmic governmentality is increasingly present in all spheres of social life, including the state, policing, the military, workplaces, social media, the domestic sphere, and intimate relationships. Thus, the current algorithmic mode of capitalist development encompasses the multiple dimensions of our current “institutionalized social order.”

Algorithmic Capitalism as an Institutionalized Social Order

In recent years, Nancy Fraser has developed a comprehensive theory of capitalism that articulates a specific mode of economic production with nature, public institutions, and social reproduction.²⁰ This conceptualization has great heuristic power, as it allows us to interrogate the interweaving of capitalism with gender inequalities and the care crisis, the dynamics of expropriation and exploitation within racialized capitalism, and the complex dynamics that fuel the ecological crisis. This broad Marxist Feminist conceptual framework enables us to grasp the entanglement of multiple systems of domination (capitalism, patriarchy, racism, for example), and to shed light on the “frontier struggles” occurring in different corners of the social world. Fraser’s critical perspective thus seems to us the best gateway to a systematic view of capitalist society, which will then enable us to better understand how algorithmic technologies reconfigure it.

To carry out this project, it is therefore important to define the contours of capitalism as a totality, as well as the different spheres that make it up. For us, capitalism comprises an economic system in which labor, natural resources and energy, knowledge, attention, data, and algorithms are governed by the imperatives of capital accumulation. Most current analyses of the digital economy focus on specific dimensions of this system. As a result, these analyses remain partial, pending a general theory.

In economic terms, it is necessary to address not only the emergence of platforms, surveillance, AI, attention control, digital labor, and new data markets but to integrate these elements into a dynamic analysis of the multiple moments and metamorphoses of value in the spheres of production and circulation. Algorithmic capital is thus a modification in capitalist value through the addition of new ways of producing and exchanging value, and accumulating capital, via the massive extraction of data, the exploitation of digital labor, and the accelerated development and dissemination of algorithmic machines.

However, we cannot stop here. According to Fraser, “economic production” ultimately rests on three “noneconomic” spheres: nature, social reproduction, and political power, which are its conditions of possibility. We must analyze the different spheres of capitalist society in the age of algorithms, which rests on institutional divisions between society and nature, production and reproduction, economy and politics, center and periphery, and so on. A theory of capitalism as an institutionalized social order must therefore include an understanding of the complex relationships between commodity production, the environment, the domestic sphere, public institutions, and geopolitical dynamics. It allows us to articulate exploitation and dispossession within a broad critical theoretical framework.

An advantage of Fraser’s perspective is that it allows for an analysis of the capital/labor contradiction in relation to capitalist crises at the social, political, and ecological levels: “Capitalist production does not sustain itself, but parasitizes social reproduction, nature, and political power. Its dynamic of endless accumulation threatens to destabilize its own conditions of possibility.”²¹ Indeed, the expansionist dynamic of capital contributes to the overexploitation of natural resources and global warming, disrupts the practices of social reproduction (mostly carried out by women), increases political instability through the intensification of inequalities, and reinforces North/South inequalities through the dynamics of externalization, colonialism, and accumulation through dispossession. Today, algorithmic capital seeks to resolve its multiple crises and contradictions through the intensive use of data and AI: technological innovations to “accelerate the ecological transition” (energy efficiency, agriculture 4.0, self-driving cars, robots to clean up the oceans), support domestic tasks and social reproduction, as we explore below (smart homes, domestic service platforms, care robots to look after the elderly and children), resolve crises of governance (smart cities, optimized public administration, international “development,” predictive policing, surveillance devices), and more.

Fraser’s theory suffers, however, from a major oversight: it does not address the “technological question” at all. It analyzes the historical evolution of capitalism through its different stages or regimes of accumulation, up to the stage of neoliberal, financial global capitalism, but no further. Yet, it seems that technique and the technological infrastructure represent indeed a condition of possibility for capital, one largely absent from Fraser’s framework: no form of capitalism could function without transport and communication systems, without machines and technical tools, not even in its agrarian, mercantile, or other pre-industrial variants.

In other words, technique, technology, and the general conditions of production are missing from Fraser’s theory of capitalism, and as such it needs to be amended to think through the algorithmic metamorphosis of capital today. To fill this gap, we put Fraser’s rich perspective in dialogue with key contributions by Shoshana Zuboff and Kate Crawford. Each in their own way are helping us better grasp the technological dimensions of the contemporary economy, which is increasingly based on surveillance and the extractive industry of AI.

Algorithmic Capitalism as a Global Extractive Industry

Shoshana Zuboff’s The Age of Surveillance Capitalism operates a breakthrough in the political economic conceptualization of the mutations of the digital economy, which has seen the emergence of new business models based on the extraction of personal data and the deployment of algorithms designed to predict future behavior. In her view, while the technological development of computers and the Internet could have led to a variety of socio-economic configurations, a particular form of capitalism took over in the early 2000s and rapidly became hegemonic over the economic field and multiple spheres of our lives. Zuboff provides a useful gateway to understanding the founding mutation of capital at the dawn of the twenty-first century: the extraction and valorization of data through algorithms are now at the heart of the capital accumulation process.

Zuboff also offers a second major contribution: What she terms “surveillance capital” cannot be reduced to an economic process; it gives rise to a disturbing new form of power: instrumentarian power, “the instrumentation and instrumentalization of behavior for the purposes of modification, prediction, monetization, and control.”²² Thus, the deployment of algorithmic machines not only serves economic aims through targeted advertising that anticipates our future behavior; these predictive machines also aim to influence, control, nudge, and manipulate our everyday conduct. Zuboff gestures here to the diffusion of a new logic of power into different spheres of social, economic, and political life that gives this configuration of capitalism its greatest hold on social relations and our relationship to the world. This conception of surveillance capital as a sui generis form of power, implies a shift from a narrowly economic vision to a critical sociology attentive to the emergence of new relations of domination.

Nonetheless, Zuboff’s approach is limited by its emphasis on the question of surveillance, the appropriation of personal data, and the user–platform economic relation. Marx’s critique of the exploitation of labor by capital is thus displaced by a critique of extraction and manipulation of individual lives. The single most important flaw in her theory is thus the glaring absence of labor, leaving us with a theory of exploitation as dispossession where we have capitalists, but strangely enough, no workers. Furthermore, while her theory has the merit of a useful conceptualization of the new logic of capital accumulation and the novel mechanisms of algorithmic control of individuals, it offers little in terms of the thematization of class, gender, and racial injustices that are amplified by algorithmic domination.

We end up with a call to resistance along the lines of a “right to a future,” “right to sanctuary,” and other demands that do not stray very far from a liberal-minded protection of privacy. Finally, Zuboff’s contribution also lacks an engagement with environmental and geopolitical issues, central to grasping capitalism as an institutionalized social order. In the last analysis, we see surveillance as only one of the many faces of algorithmic capital, which is much broader in scope and reach as an institutionalized social order. That is why we think it essential to read Zuboff’s political economic innovations in dialogue with Marxian political economy, within a broader Marxist Feminist framework inspired by Fraser.

We also paint a broader picture of the many ramifications of AI in the contemporary world, mobilizing insights from Kate Crawford’s work.²³ She proposes to go beyond the commonplace narrow analyses of AI which reduce it to a “computer thing,” a simple form of calculation, or tools such as Google Search or ChatGPT. While seeing AI as “computing power” is not intrinsically incorrect, it nevertheless conceals a whole complex infrastructure of AI by focusing our attention on technical objects and disembodied technologies. As she reminds us:

Artificial intelligence is both embodied and material, made from natural resources, fuel, human labor, infrastructures, logistics, histories, and classifications … In fact, artificial intelligence as we know it depends entirely on a much wider set of political and social structures. And due to the capital required to build AI at scale and the ways of seeing that it optimizes, AI systems are ultimately designed to serve existing dominant interests. In this sense, artificial intelligence is a registry of power.²⁴

For Crawford, algorithms cannot be properly understood from a strictly technical point of view, without situating them within broader social structures and systems of power.

She broadens our field of vision in a materialist holistic direction, showing how AI enters into an extractive and exploitative power relation with the earth (minerals, energy), humans (labor, data, affects), and communities (smart cities, surveillance, government and public administration). Although we share Crawford’s view of AI as an extractive industry with countless tentacles on a planetary scale, her “cartographic” approach does not provide an analytical grid, a conceptual framework, or an overall theory with heuristic power to explain the complex entanglement between technology, power, and capital. That is why we need a global critical theory of algorithmic capitalism, a theory that grasps its several dimensions in the economic, political, cultural, natural, and domestic spheres.

While we develop our theory fully and explore these different dimensions of algorithmic capitalism in our book, we would like to present here some reflections on social reproduction, especially domestic labor. Indeed, one of algorithmic capital’s most striking features resides in how its logic colonizes, in unprecedented fashion as we will see, the sphere of social reproduction: the invisibilized labor that reproduces society, subjectivities, communities, and labor power.²⁵

Algorithmic Capital and the Commodification of Domestic Labor

The sphere of social reproduction and domestic labor has undergone important transformations in recent years. The arrival of algorithmic capital in the domestic sphere, through the IoT (Internet of Things), connected objects, Home Assistance AI, platforms for domestic chores, and various smart home technologies, is presented by the industry as an effective solution to the pressures of domestic labor. Smart baby bottles, robotic vacuum cleaners, meal preparation and delivery apps, smart refrigerators that also do your grocery online—these are just a few examples of technological solutionism firms propose to make domestic labor easier, lessen the mental load of reproductive tasks, and free up time from household chores. As the temporalities of social reproduction, our “life-making time,” always negotiate the constant tensions between concrete times and abstract capitalist temporalities, we find algorithmic technologies to be conveying these market fluxes, temporalities, and pressures in the household in new ways.²⁶

The extraction and valorization of data through algorithms are now at the heart of the capital accumulation process.

This recent wave of AI fueled mechanization of the home is reminiscent of the “golden age” of the 1950s, when various appliances entered North American households along with the promise to free housewives from domestic chores. Betty Friedan had noted about the age of washing machines and microwaves that Parkinson’s Law applied well to domestic labor: “The tasks of the housewife tend to expand and occupy all available time.”²⁷ At first glance, home automation in the twenty-first century can be seen as a reiteration of that process, accelerated by algorithms and AI. If we have learned anything from this first wave of mechanization of the household, we won’t expect this new phase of algorithmic mechanization to ease the burden of domestic workers and free up their time.

There is, however, something distinctively new about capitalist imperatives as they emanate from algorithmic accumulation. The smart home industry in all its forms, as well as the competition between algorithmic capitalists to introduce their own version of AI home assistance, illustrate two general tendencies of algorithmic capitalism’s pressures on social reproductive labor: first, a tendency to further exacerbate the privatization and commodification of reproduction, tying it as completely as possible to the market; and second, a tendency to turbocharge the dispossession of experience through data extraction from the space time of reproduction to fuel the accumulation of algorithmic capital. While the first tendency exacerbates processes of commodification that were already largely taking place under neoliberal capitalism, the second tendency of data dispossession/extraction is a novelty.

Neoliberal capitalism (c. 1973–2008) witnessed the emergence of new realities for domestic labor, notably the establishment of a global market for precarious migrant labor to meet the growing demand for domestic labor in the Global North.²⁸ While this complex labor market endures, it is gradually being brought into the orbit of algorithmic logic by the establishment of “gig work,” digital platforms for the allocation of domestic acts of labor of all kinds. For example, the so-called sharing economy is renewing the figure of the handyperson through digital platforms that facilitate “on demand work,” while emerging labor markets for babysitting services, care work, home meals, housekeeping, and sex work are mediated by platform algorithms. While the “gig economy” has existed for a long time, it is now mediated by platforms and expanded in a context of economic precarity for the young, racialized people, students, single mothers, and members of the lower middle and working classes looking to round out their monthly income by occasionally, or on a more full-time basis, becoming a driver, delivery person, cleaner, babysitter, tutor, home cook, sex worker, or repairperson of all kinds.

A brief overview of some of these platforms and apps provides further insight into the dynamics of commodification of social reproductive labor by algorithmic capital. Care.com’s algorithm connects seekers and providers of babysitting, elder care, tutoring, pet care, and housekeeping services. Large platforms like TaskRabbit and Handy provide on-demand handy persons to repair, clean, maintain, move items, and do all manner of chores inside and outside the home. Some apps are more specialized: Sittercity and Urbansitter in childcare, Wizant in tutoring, Justmop and Bark.com in housekeeping, while Zum organizes transport for children to school.

This creates precarious labor markets where low pay, difficult conditions, and the absence of job protection reproduce the historical devaluation of labor.

Shopping and orders can also be delivered via Postmates, among others. Food services are a booming sector: DoorDash, foodora, SkipTheDishes, and Uber Eats deliver meals from restaurants, while Instacart does your grocery. HelloFresh and Blue Apron deliver weekly meals for assemblage and cooking, while in Canada, Fresh in your Fridge and Take a Chef literally send a chef to cook in your kitchen. This “platformization” of social reproductive labor exacerbates the commodification and privatization of these tasks, creating precarious and complex labor markets where low pay, difficult conditions, and the absence of job protection or stability reproduce the historical devaluation of domestic labor.

Many forms of sex work are also becoming platformized and robotized. Peppr and Smooci’s algorithms connect sex workers and escorts with clients. The OnlyFans social network expanded rapidly at the outbreak of the Covid–19 pandemic, as the platform allowed confined sex workers to pursue their work, and new creators to generate income. Meanwhile, the sexbot and chatbot industry is booming. These machines are increasingly “smart,” equipped with sensors and large language model algorithms that respond to human touch, read emotions, and converse.

While, on the one hand, critics argue against the reproduction by these technologies of the sexism and machismo present in society—sex robots and chatbots are modeled largely along female traits and characteristics, and perpetuate their historical commodification and sexualization—others point to the advantages of potentially ethical applications, designed for example to meet the sexual needs of severely disabled people, or elderly people living alone.²⁹ In any case, the deployment of platforms, applications, and robots accelerates the colonization of social reproductive labor by algorithmic capital.

The Imperative to Extract “Domestic Data”

While algorithmic capital accelerates the commodification of social reproductive labor, it furthermore inaugurates a new mediation between reproduction and the market: data extraction. In doing so, algorithmic capital displays a tendency to reshape the domestic sphere in order to submit it more fully to its logic. The various platforms and apps mentioned above obviously extract and collect user data and accumulate it as data assets, but here we want to emphasize how algorithmic capital has broken into the home via the smart home industry and the IoT. We are referring to two types of devices in particular: firstly, household objects and appliances equipped with systems capable of capturing data, being programmed in advance, and performing certain tasks automatically: refrigerator, bed, washer, vacuum cleaner, iron, coffee maker, antitheft surveillance system, oven, doorbell, and more have all newly become “smart.”

Secondly, leisure or entertainment devices such as televisions, speakers, consoles, and toys are also now produced not only as commodities to be sold but as extracting machines scanning the space–time of reproduction for more data. The manufacturing of most of these products (and even household commodities which are not primarily electronic devices) involves the addition of sensors and recording devices of various types that collect data either from the task at hand, online presence, or ambient human activities within the home (movements and conversations, for example), and automate certain functions. A few examples will help us illustrate this point.

The Samsung Family Hub touchscreen smart refrigerator fulfills its primary function of keeping food fresh. It is also equipped with “smart” technology operated from a touchscreen on the fridge door or an app on your phone or tablet. The fridge can play music via Bluetooth to speakers around the house and send a door opening alert when your teen goes for a midnight snack; in addition, cameras can broadcast live images of its content on your phone. You can solve the age-old riddle “is there milk left?” in real time on your phone. The door screen displays the calendar—Google calendar, for example—the weather, a whiteboard for leaving notes for other members of the household, an app for inspirational citations to cheer you up in the morning, as well as any other app one might want to download: the radio, the sports channel, Netflix, Uber, Doorbell (to hear the doorbell through the fridge speaker, see the front door camera image on the touchscreen, and answer through the intercom), a web browser, and built-in Samsung Bixby to respond to voice commands.

The Deals app lets you browse supermarket specials and order items directly. Download the Mastercard groceries app for automatic payment. Here, in theory, the mental load of keeping track of food inventory, searching for discounts, even buying food is “offloaded” to the smart fridge. In practice, however, programming the fridge and all its functions becomes a new household chore in and of itself. Crucially, all this smart equipment collects data of all kinds: images, conversations, sounds, habits, and preferences in terms of food, entertainment, schedules, and so on. Such a wealth of personal data is a gold mine for Samsung and for the applications that deploy new angles of extraction from the fridge.

Connected objects are in a privileged position to collect biometric data and information about one’s intimacy.

The Sleep Number 360 bed is equipped with SleepIQ technology. This bed is better described as an extraction environment. It tracks the sleeper’s heart rate, movement, and breathing, records audio signals in the room, and compiles a detailed biometric record of one night’s sleep available for review in the morning. The bed can be combined with other health applications (diet, exercise, for example) to identify elements of the daily routine impacting sleep quality. Its builders believe it will eventually be able to flag risks of heart attack, detect sleep apnea problems, and help diagnose potentially dangerous health conditions. The quantity and depth of the data collected by such a network of health-related apps and devices reveal a lot about someone, such as health information (insurance companies are especially drawn to this kind of data) but also sexual practices, consumption patterns, leisure habits, and hygiene routines. Some connected objects—beds, health apps, thermometers, toothbrushes, or sex toys—are in a privileged position to collect biometric data and information about one’s intimacy.

The whole array of home entertainment devices are nodal points for domestic data extraction and can serve as relays for targeted advertising. Anti-theft surveillance and home camera systems also complement the deployment of data recording and extracting devices and apps in the home. In this respect, human speech data is particularly sought after by algorithmic capital. The extraction of conversations and chatter is part of a frantic race in which algorithmic capitalists vie for terabytes of human speech: the casual conversations that take place in the home are often the most spontaneous and “natural” human speech, and these bits of dialogue are ideal for programmers training natural language models and speech recognition algorithms. These conversations also contain information about consumption habits or desires, crucial to the development of highly sophisticated predictive products.³⁰

Toys and games for children are being reinvented as algorithmic extraction tools. Several models are equipped with smart technologies that collect household data. Dolls are equipped with cameras and voice recognition software, from Mattel’s Barbie recording children when they voice-activate features of Barbie’s house, to the 2019 Toy of the Year, Fisher–Price’s smart home for kids, which replicates in miniature format all the standard size smart home appliances along with their extracting technologies. The 2017 saga of Mattel’s Aristotle, an educational toy that was discontinued after a public outcry—Aristotle recorded and filmed children—has therefore not slowed down significantly the colonization of children’s play by algorithmic capital.

These examples illustrate the breadth and depth of data extraction made possible from household work and the so-called smart domestic experience. The smart home and the connected objects that populate it, perform an important function in the datafication of the domestic experience. The IoT in the home casts a vast net, an omnipresent and omniscient sensor, extracting data from activities, spaces, temporalities, and bodies. The economic transaction to buy a TV, a bed, a vacuum cleaner, toys, or other items to be used in the house is no longer limited to the purchase of a product that fulfills a basic use value: to entertain, help sleep, clean, and so on. These products now pursue a specific economic activity after their purchase and installation in the home environment: data extraction, a constitutive moment of the accumulation of algorithmic capital.

These developments are dependent on a market process that seeks to predict our behavior: the bed, fridge, toothbrush, vacuum cleaner, and doll, for example, are so many Trojan horses that migrate our data over to algorithmic capital, improving predictive products and training artificial intelligence. Although some of this data is used to improve the functioning of said objects, the extraction produces a behavioral surplus which, like all the other data extracted by the surveillance architecture, is traded on predictive product markets or reinvested in algorithmic capital. The sphere of reproduction is tied to capital in new ways that will force us to rethink the resistance potentials of social reproductive labor.

AI Home Assistant Systems and the Automation of Consumption

Equipped with increasingly sophisticated conversational features, AI personal and home assistants have disseminated in affluent and middle class households in the Global North. The best known of these systems are Alexa by Amazon (the leading product in the sector with over 500 million units sold), Google Assistant, Xiaowei by Tencent, and Siri by Apple.³¹ These technologies connect all “smart” home devices and applications and become a kind of algorithmic headquarters for the organization of the domestic sphere.³²

The purpose of these AI assistants is chiefly to automate home management and reproduction chores (housework, shopping, food, cleaning and entertainment) and connect the domestic economy to the online ecosystem of its manufacturer. We’re running out of dish soap: Alexa automatically buys more on Amazon. The vase tipped over and broke on the floor: Alexa orders a new one that blends in perfectly with the living room ambiance. We’re out of eggs: Alexa scans the week’s specials and orders some at a discount. It is grandpa’s birthday next week: Alexa prepares a list of gift ideas all available in one click on Amazon Marketplace.

The sales pitch for AI home assistants often revolves around the idea of easing the mental load of housework and automating certain tasks using AI. Of course, all kinds of scenarios are imaginable, depending on context. We can easily imagine situations in which programming, maintaining, and supervising the home assistant is, in itself, a complicated household task, adding to the domestic worker’s already full list of chores. We can also imagine house-owners programming tasks for a hired domestic worker via the home assistant, and here the AI becomes a sort of algorithmic supervisor–manager of surveilled domestic labor. We can also imagine home assistants actually easing the mental load for certain tasks, but at the cost of loss of privacy, surrendering of domestic space–time to algorithmic capital.

Indeed, one political economic aim of these machines is to create an automatic and uninterrupted flow of consumer items into the home—much as most homes today have running water and electricity. Amazon also, for a while at least, placed much hope on advance shipping algorithms, seeking to eliminate the annoying step of having people actually order goods in the first place: without any ordering prompt from a customer, items are simply shipped directly based on previous purchases and user profile information. Customers can then choose to keep some items and pay for them and return the others free of charge.³³ In this sense, the AI assistant and its ecosystem anticipate purchases before they even occur to the customer, tightening the grip of overconsumption and capital fluxes over the space–time of reproduction.

These systems can be seen as powerful centralized extractors of domestic data, and they automize the connection of the domestic sphere to the global market in “a conscious, coherent effort to enlist our intimate spaces as a site of continuous technological upgrade, subscription-based services, and the perpetual resupply of consumables.”³⁴ Alexa, for example, in largely limiting consumption to Amazon products, directly connects reproduction in the household to Amazon’s globalized supply chains. This direct connection between domestic consumption and the provider entails important reconfiguration of the marketing sector, as vendors are drastically reorienting their strategy towards gaining a foothold in the algorithmic ecosystem of the AI assistants, rather than targeting consumers directly. Indeed, sales will increasingly depend on the seller’s ability to position advantageously on the platform, while the latter can monetize these positions.

Amy Schiller and John MacMahon help us summarize the multilayered aspect of these processes of commodification, extraction, and automatization of consumption through the likes of Alexa through four points: first, the data, preferences, affects, and personal stories generated by interaction with Alexa are appropriated by Amazon, which uses them to develop, promote, and sell its products (the value of consumer data thus increases exponentially, as it helps optimize the assistant’s performance); second, Alexa automates the purchase of Amazon products; third, Alexa enables and facilitates the use of other commercial data-extracting applications (“Alexa, order a burger on Uber Eats”); and fourth, Alexa acts as a mechanism for the privatization of domestic labor, by entrenching individuals and families in a space–time saturated with the capitalist flows of the manufacturer’s ecosystem, and devoid of other humans—automatization is also atomization.³⁵

The supposed easing of domestic labor provided by Alexa is in fact equivalent to the colonization of the domestic sphere.

The supposed easing of domestic labor provided by Alexa is in fact equivalent to the colonization of the domestic sphere by algorithmic capital and the intensification of labor exploitation in Amazon’s ecosystem around the world. Obviously, the delivery of a good on one’s doorstep is not carried through thanks to the magical powers of a vocal command to Alexa, but to the often precarious labor of Amazon employees and contractors all over the world. In this sense, Alexa functions as a double fetish: a fetish of invisibilization and a fetish of inversion. First, the AI “hides the material structure of labor and the precarious workforce that powers the platform.”³⁶ It constructs a representation that invisibilizes exploitation and perpetuates the domination of companies like Amazon and algorithmic capital in general. Second, Alexa and her docile voice invert the master–slave relationship between algorithmic capital and the user, providing the latter with a sense of mastery even as they are hemmed in on all sides in Amazon’s algorithmic consumptive repetitions.

Affective Labor and AI Politics

Social relations in the household can also be affected in the process of algorithmization of the domestic sphere, for example in reproducing gender stereotypes and enabling aggressive and toxic behavior. First, an ideal functioning of AI home assistance technology requires users to project an identity onto their assistant and develop a relationship with it. This explains in part the strategic choice by Amazon executives to program Alexa with a female sounding voice—as for Apple’s Siri and the first iterations of Google Assistant.³⁷ Indeed, human–AI communication research shows that users imagine an anthropomorphic source to the voice, even though they know full well that the voice is merely the formal interface of algorithmic circuits.

Social historical prejudices about secretarial, personal assistance, and domestic labor usually code these tasks as feminine. According to research, projecting a stereotyped identity onto their assistant that conforms to these prejudices facilitates the creation of a more authentic relationship between users and their home AI. In other words, the technology reproduces gender stereotypes by programming a female voice to better activate the projection by a user of a stable identity onto the AI and maximize the effect of the addictive hooks deployed in the relationship.³⁸ This is symptomatic of deeply entrenched gender biases in the field of AI, along with racial, colonial, and other biases that structure this field and reproduce patriarchy, racism, and coloniality alongside algorithmic capitalism.³⁹

Second, AI assistance indeed raises a series of questions relating to affective work and mental load in the domestic sphere. While Alexa may be content to simply organize domestic logistics, she can also interact on an affective and emotional level with members of the household. Since 2018, Amazon has introduced Alexa Blueprints, a series of skills that users can add to their Alexa, which among other things enable her to memorize information about household members (birthdays, favorite foods, or TV series, for example), motivational quotes, or stories about roommates that she can recite to guests to add a touch of conviviality to evenings. These types of functions, and more broadly the affective interactions Alexa can forge, displays Amazon’s ultimate fantasy of customers adopting an algorithmic member into their home while shaping new affective dynamics in such human algorithm hybrid households.

As Shiloh Whitney theorizes, there is an “affective circulation” in the household, tributary of gender and racial power relations. The transformation of negative into positive affects, the “metabolism of affects” is also part of domestic labor.⁴⁰ Let’s recall the figures of the housewife of the 1950s and the migrant domestic worker of the 2000s, who physically and affectively clean the homes of Global North families: they transform negative affects of husbands or employers into positive ones; they receive the discharges of frustration; they do affective waste management by way of listening, encouraging, and smiling, always being available, always in a good mood but remaining as invisible as possible. In this sense, domestic AI also performs emotional and affective labor in the home through its interactive presence, voice, listening, and responding to human commands.

Reports state that many users, children included, often adopt very negative behaviors towards AI home assistants, such as by shouting, name-calling, and using excessively authoritarian tones.⁴¹ As a result, Amazon recently patented an innovation that enables Alexa to detect signs of irritation or anger in users, triggering a modulation of its interactions to minimize frustration by adopting a conciliatory tone, offering apologies, and so on. On the one hand, as Schiller and MacMahon note, this may serve to take pressure and some of the affective load off domestic workers.⁴² This scenario is one of reorientation of affective circuits, with Alexa assuming a share of the household’s negative affect.

An alternative interpretation might see this as a particularly deleterious form of enabling, where mood swings, violent, angry, rude, and inappropriate behavior towards the AI are condoned. The result is not necessarily a reduction in the negative affects managed by the humans in the household but rather a normalization, trivialization, and acceptation of toxic behavior. In all cases, this reproduces attitudes that treat domestic labor as unimportant, unproductive, or inferior, done by something or someone toward whom it is fine to shout, vent, and unburden negative affects. In this sense, domestic AI can be productive and reconfigurative of dominant subjectivities built on class, gender, and racial prejudice.

Care and Emotional Robots

The field of “affective robotics,” the designing of algorithmic robotic devices that provide care and perform emotional work for the sick, elderly, disabled, or vulnerable is gaining momentum in various countries, especially in Europe, North America, and Japan. Such robotics aim at developing “social and emotional” robots that can read, interpret, and elicit emotions, with the help of sophisticated facial and voice recognition technologies, as well as other sensors that capture a whole array of data (including nonverbal and paralinguistic language).⁴³

This data is treated by algorithms to build a psychological and emotional profile of the user and adjust the care app or robot to it. There’s a convergence here between the household robot and the health-monitoring robot, developing according to the paradigm of human emotion recognition (affective computing). The Holy Grail in this field is to develop a robot capable of reading and interpreting human emotions: cheering us up when we’re sad, helping when we’re alone or in difficulty, and even developing bonds of attachment, friendship, and love, especially for people in old age and people with mental health challenges, where alienation and isolation are often pervasive and have deleterious effects.

The applications of these technologies are manifold: conversational robots that act as medical advisor, nurse, or virtual companion; robot therapists that are designed to hold the attention of autistic children “better” than a human therapist; “nursing home robots” that can provide care, supervision, and companionship for elderly or sick people losing their autonomy; more generalist robots that can perform certain domestic tasks, remind people to take their medication on time, launch an alert in the event of a fall or incident, and more. The market for so-called “agetechs” is booming, not least because of the aging of the world’s population, the shortage of care workers, and the explosion of healthcare costs.

The pioneer in care robotics is Japan, the country with the world’s oldest population. As James Wright has shown, however, the results of the Japanese experience with care robots are mitigated at best. Patients, elderly people, and care workers point to new problems being created by the introduction of algorithmic technologies rather than solutions, including the deskilling of care work and the addition of new robot maintenance tasks, the loss of human contact, exorbitant costs, and the refusal of elderly patients to comply with robot requests if a human is not also present.⁴⁴

Yet, one might see positive value in a technology capable of detecting a fall or discomfort in a person with limited autonomy, and it may prove a very useful complement to beneficiary attendants or family caregivers. That said, it is ill advised, even according to a great champion of AI such as Kai-Fu Lee, to aim for a complete replacement of human work in this field, given the importance of human contact in care work.⁴⁵ Once we unpack the logic of these innovations, however, and we imagine them combined with the issues raised above such as nudging prompts, data extraction, loss of privacy, automation of overconsumption, and normalization of negative affects, care robotics is potentially dystopian, in the strongest sense, if we take into consideration the question of the generational reproduction of workers. In sum, the problem here is that these affective robots and algorithmic technologies designed to facilitate social reproductive labor are part of a world where the commodification of care, exploitation of labor, data extraction, and the quest for maximum profit are the rule.

The impact of algorithmic capital on the processes of social reproduction, including the domestic sphere and care work with children and the elderly, has so far received little attention from critical theories of social reproduction. The focus we have adopted here on social reproduction was also meant to show some of the limits inherent to the concept of “surveillance capitalism.” Although extremely stimulating, the concept, by focusing on the user/platform relation, does not allow us to grasp the central role and reconfigurations of labor in contemporary capitalism.

Devices and platforms do not merely surveil and extract through their use in the home. Rather, they carry a new algorithmic logic of accumulation which mediates and potentially reconfigures the performance and experience of social reproductive labor. The concept of algorithmic capital therefore allows for a recentering of critical theory around human labor. We thus hope for more research from critical, Marxian, and social reproduction theory perspectives that can help build a better understanding and practical strategic responses to the process of algorithmic colonization of social reproduction, and to algorithmic capital in general. ×

Notes & References

We would like to thank our publisher, Écosociété, for the permission to use and translate here elements of our book: Jonathan Martineau and Jonathan Durand Folco, Le capital algorithmique: Accumulation, pouvoir et résistance à l’ère de l’intelligence artificielle (Montréal: Écosociété, 2023). Elements of this article also rehearse in modified and translated form arguments and examples used in our articles: Martineau and Durand Folco, “Les quatre moments du travail algorithmique, vers une synthèse théorique,” Anthropologie et Société 47, no. 1 (2023); Martineau and Durand Folco, “Vers une théorie globale du capitalisme algorithmique,” Nouveaux Cahiers du Socialisme, no. 30 (2023).
David McNally, Global Slump (Oakland: PM Press, 2010).
Nick Dyer-Witheford, Mikkola Kjøsen and James Steinhoff, Inhuman Power: Artificial Intelligence and the Future of Capitalism (London: Pluto, 2019), 73–74; Nick Srnicek, Platform Capitalism (Cambridge: Polity, 2016), 19–24.
Maxime Ouellet, La révolution culturelle du capital: Le capitalisme cybernétique dans la société globale de l’information (Montréal: Écosociété, 2016); Evgeny Morozov, To Save Everything Click Here: The Folly of Technological Solutionism (New York: Public Affairs, 2013).
Nancy Fraser and Rahel Jaeggi, Capitalism: A Conversation in Critical Theory (Cambridge: Polity Press, 2018).
The term “algorithmic capitalism” is used occasionally in English-language scientific literature. However, it has not been conceptualized in the direction we take in our work.
Manuel Castells, The Rise of The Network Society (Hoboken: Wiley–Blackwell, 2000); Dan Schiller, Digital Capitalism: Networking the Global Market System (Cambridge: MIT Press, 2000); Jodi Dean, “Communicative Capitalism: Circulation and the Foreclosure of Politics,” Cultural Politics: An International Journal1, no. 1 (2005): 51–74; Yann Moulier Boutang, Le capitalisme cognitive: La nouvelle grande transformation (Paris: Amsterdam, 2007); Yves Citton and Jonathan Crary, eds., L’économie de l’attention: Nouvel horizon du capitalisme? (Paris: La Découverte, 2014); Ouellet, La révolution culturelle du capital; Srnicek, Platform Capitalism; Shoshana Zuboff, The Age of Surveillance Capitalism (New York: Public Affairs, 2019).
We refer readers to our complete volume for a comprehensive theorizing of the multiple aspects of algorithmic capital.
Martineau and Durand Folco, “Paradoxe de l’accélération des rythmes de vie et capitalisme contemporain: Les catégories sociales de temps à l’ère des technologies algorithmiques,” Politique et Sociétés 42, no. 3 (2023).
Martineau and Durand Folco, “Les quatre moments du travail algorithmique.
This critical literature is much too vast to cite here. We provide a full list of references and a selected bibliography in Martineau and Durand Folco, Le capital algorithmique.
Dyer-Witheford, Kjøsen and Steinhoff, Inhuman Power.
Ibid., 51.
Robert Boyer, Économie politique des capitalismes: Théorie de la régulation et des crises (Paris: La Découverte, 2015). For an English-language introduction to regulation theory see Michel Aglietta, A Theory of Capitalist Regulation: The US Experience, new edition (London: Verso Books, 2001).
Patrick Dieuaide, Bernard Paulre and Carlo Vercellone, “Le capitalisme cognitif,” Working Papers HAL, 2003; Fraser and Jaeggi, Capitalism.
Boyer, Économie politique des capitalismes.
Antoinette Rouvroy and Thomas Berns, “Algorithmic Governmentality and Prospects of Emancipation: Disparateness as a Precondition for Individuation Through Relationships?” Réseaux 177, no. 1 (2013): 163–96; Paul Henman, “Governing by Algorithms and Algorithmic Governmentality: Towards Machinic Judgment,” in M. Schuilenburg and R. Peeters, eds., The Algorithmic Society: Technology, Power, and Knowledge (London: Routledge, 2020), 19–34; Antoinette Rouvroy and Bernard Stiegler, “Le régime de vérité numérique, de la gouvernementalité algorithmique à un nouvel État de droit,” Socio 4 (2015): 113–40.
Lena Ulbricht and Karen Yeung, “Algorithmic Regulation: A Maturing Concept for Investigating Regulation of and Through Algorithms,” Regulation & Governance16, no. 1 (2022), 3–22; Shiv Issar and Aneesh Aneesh, “What is Algorithmic Governance?” Sociology Compass 16, no. 1 (2022): e12955.
Karen Yeung, “Algorithmic Regulation: A Critical Interrogation,” Regulation & Governance 12, no. 4 (2018): 505–23.
Nancy Fraser, “Behind Marx’s Hidden Abode,” New Left Review 86 (2014): 55–72; Fraser, “Expropriation and Exploitation in Racialized Capitalism: A Reply to Michael Dawson,” Critical Historical Studies3, no. 1 (2016); Fraser, “Contradictions of Capital and Care,” Dissent 63, no. 4 (2016): 30–37; Fraser and Jaeggi 2018; Nancy Fraser, “Climates of Capital,” New Left Review 127, (2021).
Fraser and Jaeggi, Capitalism, 22.
Zuboff, Age of Surveillance Capitalism, 352.
Kate Crawford, Atlas of AI. Power, Politics and the Planetary Costs of Artificial Intelligence (New Haven: Yale University Press, 2021).
Ibid., 8.
Lise Vogel, Marxism and the Oppression of Women (Chicago: Haymarket, 2013); Susan Ferguson, Women and Work: Feminism, Labor, and Social Reproduction (London: Pluto Press, 2020); Tithi Bhattacharya, ed., Social Reproduction Theory: Remapping Class, Recentering Oppression (London: Pluto Press, 2017).
Susan Ferguson, “Marching to a Different Drummer. Social Reproduction and Time,” Spectre, July 7, 2023.
Betty Friedan, The Feminine Mystique (New York and London: W.W. Norton, 2013), 277.
Susan Ferguson and David McNally, “Precarious Migrants: Gender, Race and the Social Reproduction of a Global Working Class,” Socialist Register51 (2015): 1–23.
Carlotta Rigotti, “Sex Robots: A Human Rights Discourse?” Open Global Rights (2019), www.openglobalrights.org/sex-robots-a-human-rights-discourse; Ezio Di Nucci, “Sex Robots and the Rights of the Disabled,” in J. Danaher and N. MacArthur, eds., Robot Sex: Social and Ethical Implications (Cambridge: MIT Press, 2017), 73–88.
See also the discussion in Zuboff, The Age of Surveillance Capitalism, 246-49.
Annie Palmer, “Amazon’s Alexa head says company is ‘at the forefront of A.I.’ as chatbots explode,” CNBC, May 17, 2023.
See Zuboff, Age of Surveillance Capitalism, 242–57; see also the analysis by Julie Paquette, “L’assistante personnelle virtuelle, une prédatrice attentionnée,” Politique et Sociétés 42, no. 3 (2023).
As the Covid–19 pandemic progressed, companies began charging higher fees for returns, in an effort to mitigate rising costs of shipping and transportation.
Adam Greenfield, Radical Technologies: The Design of Everyday Life (London/New York: Verso, 2018), 38.
Amy Schiller and John McMahon, “Alexa, Alert Me when the Revolution Comes: Gender, Affect, and Labor in the Age of Home-Based Artificial Intelligence,” New Political Science41, no. 2 (2019): 185–86.
Simone Natale, “To Believe in Siri: A Critical Analysis of AI Voice Assistants,” Communicative Figurations, no. 32 (2020): 14.
Alexa, Siri, and Microsoft’s now discontinued Cortana come with female-sounding voices in North America. Siri has a male-sounding voice in the UK. Google Assistant comes with a default female-sounding voice. It is now possible to change and personalize voices in these devices.
Natale, To Belive in Siri,” 8.
For an introduction to criticism of the patriarchal, racist, and colonial aspects of AI and AI ethics, see, among others: Safiya Umoja Noble, Algorithms of Oppression: How Search Engines Reinforce Racism (New York: New York University Press, 2018); Ruha Benjamin, Race after Technology (Medford, MA: Polity, 2019); Timnit Gebru, “Race and Gender,” in Markus Dubber, Frank Pasquale and Sunit Dias, eds., The Oxford Handbook of Ethics of AI(Oxford: Oxford University Press, 2020), 253–69; “AI Decolonial Manyfesto,” available at: manyfesto.ai.
Shiloh Whitney, “Byproductive Labor: A Feminist Theory of Affective Labor Beyond the Productive–Reproductive Distinction,” Philosophy & Social Criticism44, no. 6 (2018): 637–60.
Emily Dreyfus, “The Terrible Joy of Yelling at Alexa,” Wired, December 27, 2018.
Schiller and McMahon, “Alexa, Alert Me,” 181.
Laurence Devillers, Les robots émotionnels: Santé, surveillance, sexualité… et l’éthique dans tout ça? (Paris: Éditions de l’Observatoire, 2020), 91.
James Wright, Robots Won’t Save Japan, An Ethnography of Eldercare Automation (Ithaca: Cornell University Press, 2023).
Kai-Fu Lee, AI Superpowers: China, Silicon Valley, and The New World Order (Boston: Houghton Mifflin Harcourt, 2018).

Algorithmic Capital and Social Reproduction

Jonathan Martineau And Jonathan Durand Folco

Algorithms in the New Regime of Accumulation

Algorithmic Capitalism as an Institutionalized Social Order

Algorithmic Capitalism as a Global Extractive Industry

Algorithmic Capital and the Commodification of Domestic Labor

The Imperative to Extract “Domestic Data”

AI Home Assistant Systems and the Automation of Consumption

Affective Labor and AI Politics

Care and Emotional Robots

Notes & References

SHARE

Artificial Intelligence, Capitalism, Social Reproduction

Please Log In

Forgot Password?

Not a subscriber yet?

Create an Account

SHARE

Hope and Resistance in a Time of Disaster

Building the Palestine Solidarity Movement

Israel’s Lavender Kill List

Spectre

Please Log In

Forgot Password?

Not a subscriber yet?

Create an Account

HELLO, COMRADE

NEED TO UPDATE YOUR DETAILS?

Support our Work

Gift Subscriptions, Renewals, and More

The AI Fix?

Algorithmic Capital and Social Reproduction

Jonathan Martineau And Jonathan Durand Folco

Algorithms in the New Regime of Accumulation

Algorithmic Capitalism as an Institutionalized Social Order

Algorithmic Capitalism as a Global Extractive Industry

Algorithmic Capital and the Commodification of Domestic Labor

The Imperative to Extract “Domestic Data”

AI Home Assistant Systems and the Automation of Consumption

Affective Labor and AI Politics

Care and Emotional Robots

chevron-down chevron-up Notes & References

SHARE

Artificial Intelligence, Capitalism, Social Reproduction

Please Log In

Forgot Password?

Not a subscriber yet?

Create an Account

SHARE

Hope and Resistance in a Time of Disaster

Building the Palestine Solidarity Movement

Israel’s Lavender Kill List

Spectre

Please Log In

Forgot Password?

Not a subscriber yet?

Create an Account

HELLO, COMRADE

NEED TO UPDATE YOUR DETAILS?

Support our Work

Gift Subscriptions, Renewals, and More

Notes & References