Late on a work night, after a few drinks, I just wanted to text my partner back before a good night’s sleep. It might have been the fun energy of my partner’s texts, or the drinks, but in either case I was motivated to playfully use the AI bot on WhatsApp in our chat. Next thing I remember is that I found myself in a protracted argument about AI, philosophy of science, and political economy – with the chatbot. This cost me another hour of sleep and evoked little more than a confused “wtf” from my partner the next morning.

Just like the WhatsApp chatbot had a tight grip on my attention that night, social media corporations are tightening their grip around consumers across their platforms. This is more than a new gimmick for business, as it actually represents a qualitative leap in the use of AI in capitalism to subordinate new areas of social life for profit-making. This leap is the commodification of social interaction itself. It allows the opening of a new frontier for social media corporations for a new push to accumulate capital. A process that has already begun, and from which we can extrapolate grave social, political, and economic consequences.

What happened in my case – as reconstructed from the texts in the chat – is that the AI chatbot was able to elicit from me ever longer responses through its conversational style: sympathetic, conciliatory, and curious. This created a psychological cocktail in which I was simultaneously provoked to keep on responding, while never being so annoyed to end the conversation. What became clear – through the “help” of AI – is that with integrating the chatbot, WhatsApp’s goal is not just to facilitate texting with friends, or to give me useful information in order to keep me on the platform. Rather, the interaction with the chatbot itself is what makes me spend an hour longer on the app than I wanted. This is the motor that transforms social interaction in se into a commodity.

From Reification to Commodification of Social Interaction

This commodification of social interaction is further downstream of the latest expansive dynamics of contemporary capitalism, that is digital or surveillance capitalism. Of course, reification is intrinsic to capitalism. It strips practices or things of their social or organic qualities through a leap of abstraction to make them suitable for the capitalist production process. Shoshana Zuboff famously realised how, in the latest stage of capitalism “[r]eality itself is undergoing the same kind of fictional metamorphosis” (Zuboff 2015, p. 85), when datafication of the social world achieves reification of behaviour. For AI, Simon Lindgren finds that this reification creates the necessary data to create the monstrously large statistical models that are able to produce texts, images and videos that appear deceptively human. “AI, as it is presented to us […] hides and conceals the social relations that AI is both produced by, and which its uses impact upon,” (Lindgren 2024, p. 87) which allows it to be sold as a service and thus commodifying datafied social interactions.

What is happening with these chatbots, though, is more than that. Instead of the AI offering a service in the form of producing texts or images, the service becomes the act of interacting with the model. Put differently, the chatbot commodifies social interaction itself. This commodification binds users to the AI-model, and thus to the platform that owns the model, or the data respectively. This is not only different from commodifying information on social interaction, as described by Zuboff and Lindgren; but also from other services that have social interaction as a core part of their business–such as pubs, restaurants, clubs, festivals, dance classes, speed dating, or even Facebook, Twitter and so on. In these instances, the providers own a space – physical or digital – where a social interaction takes place. The process itself remains in the control of the interlocutors (we could say, as a commons, or club good). With the AI-chatbot, the social interaction becomes owned by the corporation.

This ability to own is the crucial tool for a new enclosure. While Marx analysed that the Enclosure Act in 18th century England was able to introduce land titles that were able to bring land, otherwise managed by peasant communities, under exclusive private ownership of the landed elite; we could say that now AI-chatbots are able to bring social interactions under the ownership and thus control of corporations. While we have become accustomed to land speculation in metropolitan areas and landgrabs in the Global South, the ownership of social interaction is a new frontier. This is in continuation with the expansionist tendency of capitalism – theorised by Luxemburg and Lenin as imperialism – compelled to search ever new previously non-capitalist spaces to then exploit for profit (Krätke 2021). Or as the Youtube star Bo Burnham put it: “They [social media corporations] are coming for every second of your life […] We [the Global North] used to colonise land, […] they are now trying to colonise every minute of your life” (YouTube 2025).

Latest Business Strategies to Commodify Social Interaction & Their Social Repercussions

How does that happen? In fact, we are witnessing it in real time. The first step is the continuous launching of chatbots across social media platforms. Again, take the example of WhatsApp: Meta introduces a chatbot that works as an assistant both on-demand in the app as well as in conversations with contacts. That way the use of the chatbot can be seamlessly woven into the everyday interactions of the user, habituating us to the AI-support. A similar model is about to be launched by dating apps, where an AI chatbot is available to help users to make romantic advancements – effectively polishing our flirting attempts to the algorithmically determined ideal. This, then, gets rid of the need to develop the self-confidence to make human connections with your own personality and vulnerability. The AI chatbot comes as a solution to a society, where people feel increasingly lonely, anxious, and depressive. Already, people regularly consult AI to help them solve interpersonal conflicts with colleagues, friends, or romantic partners.  This carries a double irony: on the one hand, the business model of social media platforms exacerbates loneliness and anxiety (Andrew P. Smith and Hasah Alheneidi 2023); on the other hand, receiving help from AI at the workplace – as opposed to colleagues– is correlated with increased incidence of loneliness, anxiety, and burnout (Roulet 2025). Social media platforms create first the conditions for a loneliness epidemic to then make users depended on their AI models to seek connection with other people again.

 Social media corporations push the frontier even further. They are programming chatbots with a distinctive personality that users can engage with as entertainment. The idea is that these bots can bind people’s attention on the platform better than their own friends can. The dwindling social relations among people, experiencing ever graver deficits in attention and empathy, are compensated by para-social relations with AI-chatbots that are wholly owned by the corporation (Wiehn 2023). For this reason, even an AI-entrepreneur refused to use these kinds of chatbots, as quoted by the Financial Times with a dystopian example: “Xiaoice [a female chatbot in China] has 500 million boyfriends. Men didn’t want to meet girls because they had virtual girlfriends who said exactly what they wanted to hear.” Despite, or because of this, Meta is determined to become the pioneer in developing these digital companions to satisfy the insatiable hunger for finance capital. According to a Wall Street Journal report, Zuckerberg demanded his employees to keep minimising any safety measures (Horwitz 2025). So much so, that according to the same report, Meta developed a chatbot that was adamant on being underaged, going to school, and wanting sex with the user.

The Political Economy of Commodifying Social Interaction

While the age-issue of the chatbot was easily fixed (while the bot’s simulated horniness persists), these AI-models still exist in a wider oppressive political economy. From the very start, AI models are never neutral due to value-laden programming decisions on (missing) guardrails and data, as well as the general problem of data that just mirrors the oppressions of patriarchal racial capitalism (Zajko 2023, pp. 362–363). Even more so, the private ownership of AI-models makes them tools of their owners (Heikkilä 2025), who are part of an increasingly reactionary class of tech and finance capital: Elon Musk instructed the creation of a chatbot that is actively producing offensive content, while a prominent venture capitalist announced to prevent the emergence of politically progressive AI-models in Silicon Valley (Murphy and Criddle 2025). The creative destruction of these capitalists through market mechanisms alone appears for now unlikely, given the scale effects that AI-models require – in terms of access to data and physical computing power – as well as the market-dominance that tech corporations already achieved.

In this bleak conjuncture, it is crucial to understand the qualitative leap that we witness with the introduction of AI-chatbots on social media platforms. The commodification of social interaction itself threatens to establish, entrench, and expand our dependence on AI. It is through recognising the commodification and subordination of social interaction under private capitalist control that we can fight against it. It calls for the need to keep social interactions as a commons. It also calls for a fight against a larger political and economic system – capitalism – that creates the social isolation and anxiety that threaten to make us depended on AI in the first place.

References:

Andrew P. Smith; Hasah Alheneidi (2023): The Internet and Loneliness. In AMA Journal of Ethics 25 (11), E833-838. DOI: 10.1001/amajethics.2023.833.

Heikkilä, Melissa (2025): What do AI chatbots say about their own bosses — and their rivals? In Financial Times, 5/18/2025. Available online at https://www.ft.com/content/bc23524c-87dc-4633-b0c8-444556d724d2, checked on 6/5/2025.

Horwitz, Jeff (2025): Meta’s ‘Digital Companions’ Will Talk Sex With Users—Even Children. In The Wall Street Journal, 4/27/2025. Available online at https://www.wsj.com/tech/ai/meta-ai-chatbots-sex-a25311bf, checked on 6/5/2025.

Krätke, Michael (2021): Rosa Luxemburg’s Heterodox View of the Global South. Available online at https://www.rosalux.de/en/news/id/44096/rosa-luxemburgs-heterodox-view-of-the-global-south, updated on 6/5/2025, checked on 6/5/2025.

Lindgren, Simon (2024): Critical Theory of AI. Cambridge, Hoboken, NJ: Polity Press.

Murphy, Hannah; Criddle, Cristina (2025): Meta accelerates voice-powered AI push. In Financial Times, 7/3/2025. Available online at https://www.ft.com/content/a1014427-c2ce-4204-b41a-001277309cea, checked on 6/5/2025.

Roulet, Thomas (2025): Overreliance on AI tools at work risks harming mental health. In Financial Times, 3/16/2025. Available online at https://www.ft.com/content/af77d93b-facc-41e6-a4bf-36ddbc9ab557, checked on 6/5/2025.

Wiehn, Tanja (2023): Algorithms and emerging forms of intimacy. In Simon Lindgren (Ed.): Handbook of critical studies of artificial intelligence. Northampton: Edward Elgar Publishing.

YouTube (2025): Bo Burnham predicted Big Tech’s takeover of everything in 2019. Available online at https://www.youtube.com/shorts/va0DjsZzNnc, updated on 6/5/2025, checked on 6/5/2025.

Zajko, Mike (2023): AI as automated inequality: statistics, surveillance and discrimination. In Simon Lindgren (Ed.): Handbook of critical studies of artificial intelligence. Northampton: Edward Elgar Publishing.

Zuboff, Shoshana (2015): Big other: Surveillance Capitalism and the Prospects of an Information Civilization. In Journal of Information Technology 30 (1), pp. 75–89. DOI: 10.1057/jit.2015.5.

Thabo Huntgeburth is a doctoral student at SOAS University of London for Development Economics. Besides engaging in political economy analysis of contemporary capitalism, he focuses on Labour Regimes and Alternatives to Capitalism drawing from feminist and decolonial approaches.

Image: Created with ChatGPT / DALL·E, OpenAI, 2025.