Sage Journals: Discover world-class research

Abstract

In this article, the main focus will be to analyze the notion of mediation in an attempt to apply it to one of the major topics of our time: the increasing impact of AI technologies in social life. Mediation being essentially a relational concept, I will start by linking it to the relational thinking developed in sociological theory and postphenomenology. My own contribution will be to take a substantialist position and to define what I consider to be the ecstatic dimension of relationality, re-interpreted via the notion of “being-toward.” I will also draw on Sartre's socio-materialist understanding of mediation, focusing in particular on his general conceptualization of the “practico-inert field.” A metaphorological approach will be presented in the final part of the study to illustrate the relations between humans and AI technologies. Three metaphors will be used to explore the meaning of algorithmic mediations: “vampire objects,” “bewitched quantities,” and “frozen voices.” These metaphors will be interpreted in light of empirical phenomena: (a) algorithmically generated “synthetic data,” (b) “high-frequency trading,” and (c) AI recruitment chatbots in hiring.

Keywords

mediation relational ontology substantialism AI technologies algorithms socio-materiality practice

Introduction

What does the concept of mediation entail at a time when new generative AI technologies are transforming organizational evaluation and decision-making practices? As an inherently relational concept, mediation refers to processes of connectivity and coordination, thus penetrating the core of the type of relational ontologies elaborated by sociologists who have sought to overcome dichotomies like structure/agency and individual/society (e.g. Bourdieu, 1998; Burkitt, 2016; Crossley, 2011; Dépelteau, 2018; Emirbayer, 1997). However, most relational sociologists tend to use assumptions formulated around a non-substantialist perspective on people, things, and reality in general. Every individual is comprehended as a socially mediated actor; the social infiltrates “every nook and cranny of the human person” (Donati & Archer, 2015, p. 7; cf. Roumbanis, 2010). Meanwhile, philosophers of technology working within the post-phenomenological paradigm and Latourian actor-network theory have also argued strongly against the idea that humans and things have any form of pre-given nature, putting instead all emphasis on mediation (Ihde, 1990; Verbeek, 2012).

In the present study, I will engage both critically and constructively with the aforementioned ideas from relational ontology by introducing an alternative framework, my main objective being to theorize the notion of algorithmic mediation. The investigation is divided into three parts, where one of the building blocks is a socio-materialist understanding of mediation. However, before I elaborate on that part, I will first establish what I consider to be a proto-phenomenological view of relationality using the concept of “being-toward” to pinpoint my own substantialist position,¹ which I wish to contrast with the dominating non-substantialist trend in social theory mentioned above. By taking this contrasting position, I am thus assuming the relative independency and discreteness of all individual beings, which can be related to what Gasché (1999, p. 10) correctly noted: “relation has with respect to substance, and its unity, a multiplying power. It secures the difference of things, their singularity.” The notion of being-toward will be used to depict the complex multi-dyadic level of mediation, which I consider strongly linked to the intentionality of lived experience (Husserl, 1900/2001). In classical phenomenology, human consciousness is defined in relation to its directedness toward either real or imagined phenomena. Intentionality also relates to experiences of absence and nothingness in everyday life, thus adding a profound element of intricacy and tension to our general understanding of mediation. Relatedly, Scott (2022) analyzed the realm of “negative social phenomena,” which refers to everything that does not empirically exist, yet may have a meaningful impact on a person's life trajectories. The notion of being-toward will constitute the theoretical core of my ecstatic definition of relationality (from the Ancient Greek word “ekstasis”, ek-meaning “out” and stasis meaning “standing” or “position”), echoing what Simmel (1908/1971, p. 15) declared over a century ago: “the individual can never stay within a unit which he does not at the same time stay outside of, that he is not incorporated into any order without also confronting it.” From this multi-dyadic level of relational thinking, I will then turn to the second part of my theoretical investigation and search for the socio-material foundations of mediation. To be sure, this should not be interpreted as a dilution of the ecstatic dimension, but rather as an extension and amplification, thus putting the “problem of human interdependencies” at the very heart of social theory (see, Elias, 1978, p. 134).

Hence, I will elaborate on a broader trinitarian relational ontology by drawing on Jean-Paul Sartre's (1960/2004) Critique of Dialectical Reason (hereafter Critique). In his treatise, which is a rather idiosyncratic blend of phenomenology and historical materialism, Sartre introduced a theoretical motif that revolves around the pluri-dynamic relationships between the “movement of individual praxis, plurality, the organization of plurality, and the plurality of organizations.” And he added, “One need only open one's eyes to see this. Our problem concerns these connections [my italics]. If there are individuals, who, or what, totalizes?” (Sartre, 1960/2004, p. 80). This problem definition is pivotal for imagining the deep complexities, inertia, and contradictions of mediations in contemporary societies, or what Sartre used an alternative socio-materialistic name for: “the practico-inert field.” The plurality of individuals and artifacts is key to understanding the ecstatic dimension of mediations. Individuals are constantly moving between different social configurations and organizational contexts (home, workplace, public transportation, supermarkets, etc.), and there are two fundamental forms of reciprocal relations in Sartre's social philosophy: serialities and groups. One general theme used throughout the Critique to explain the notion of mediation is how human meaning and practices are constantly inscribed into matter (in books, documents, buildings, machines, large infrastructures, etc.). For Sartre, artifacts can possess important agentic properties, an idea that was later developed in a quite different direction by Latour (1994, 2005) and other actor-network theoreticians. In light of today's socio-technological developments, where generative AI is even displaying self-supervising and creative forms of actorhood, we need to extend the Sartrean understanding of mediation to new algorithmic phenomena.

In the third and final part of the present study, I will thus move from a general socio-material understanding of mediation to a metaphorological interpretation of algorithmic mediations. Here, the aim is to analyze some of the “enigmatic technologies” (Pasquale, 2015) that today organize human relations by shaping evaluation and decision-making processes. The impact of AI and deep learning models has increased exponentially during the past few years; many important opportunities and life chances—including the ability to obtain loans, employment, parole, housing, and insurance—are being automated (Beer, 2024; Bursell & Roumbanis, 2024; Cevolini & Esposito, 2020; Gerdon et al., 2022; Kahneman et al., 2021; Kiviat, 2023; Roumbanis, 2025).

I will use three metaphors to highlight key attributes of contemporary human–AI relations, my aim being to contribute to theoretical development within the sociology of algorithms and critical social theory (Amoore, 2020; Beer, 2024; Burrell & Fourcade, 2021; Howard & Borenstein, 2018; O’Neil, 2016). Using a metaphorological approach, I wish to explore the initial question I raised at the beginning of the article: What does the concept of mediation entail at a time when new generative AI technologies are transforming organizational evaluation and decision-making practices? The three metaphors I will present here are “vampire objects,” “bewitched quantities,” and “frozen voices,” all of which will be tied to contemporary empirical examples to show the emerging tensions these new types of interdependencies create in social life, and how this development, overall, is changing the practical conditions of the human being-toward.

Rethinking the notion of mediation

Substance and ecstatic relationality: being-toward what, how, and why?

The concept of relation occupies quite a prominent place in contemporary sociological and anthropological thinking (Burkitt, 2016; Dépelteau, 2018; Donati & Archer, 2015; Emirbayer, 1997; Latour, 2005; Strathern, 2018). It is perceived almost as a redeeming term that can solve all the classical problems and antinomies in social theory. Indeed, relations have gained and now possess a very special ontological status. For example, Bourdieu (1998) famously claimed that “the real is relational,” putting this statement in sharp contrast to what he called substantialist views, which see human individuals and other natural entities as embodying distinct, existentially fundamental boundaries in themselves. In line with this idea, Emirbayer (1997, p. 282) proposed that the key issue confronting all contemporary sociologists “is the choice between substantialism or relationalism.” However, I think it is a serious mistake to argue that a substantialist ontology cannot be combined with relational thinking, as doing so creates a false dichotomy and eliminates the possibility of navigating in other theoretical directions.

To approach the fundamental question of relationality at the ontological level, I will go back to the ancient Greek wording pros ti, meaning “toward something” or being-toward. Some philosophers have argued that the Latin translation of pros ti as “relatio” can be somewhat misleading, as the latter contains both subjective and social connotations that are absent in the former. That is a valid objection. Gasché (1999) explained that the word relation has been perceived as “a metaphysical reduction not only of the Greek understanding of being-toward, but also of the existential complexities of any comportment (Verhalten) and, …of the complexities of being-toward in the context of the thinking of being.” In Aristotle's metaphysics, the pros ti represents only one of 10 different categories (position, time, quality, quantity, etc.), yet he gave ontological priority to substance (ousia). Being-toward concerns how things stand in relation to each other, whereas substance regards compounds of matter and form that manifest themselves as “this something” (tode ti). At this point, what we need to do is to look more closely at the very meaning of being-toward from a sociologically relevant perspective, with a view to presenting what I interpret as the ecstatic dimension of human individuation and its multi-dyadic directedness in social life.

First of all, being-toward presupposes a place of the other, “a place that can be occupied by the opposite, or other, …but which is not saturated, fulfilled, or exhausted by this occupation” (Gasché, 1999, p. 10). This statement underscores the theoretical importance of securing the substantive difference between all things and/or living creatures—their ontological singularity. Indeed, this idea helps us visualize the depths and richness of the world and grasp the phenomenological relevance of absence, negation and “elsewhereness,” which correspond to the human experience of lack, opposition, and the distance to someone or something. These are existential dimensions running throughout human history. For example, even if the human species has changed significantly over the past three millennia, owing to socio-material, cultural, and scientific evolution (cf. Malafouris, 2013), certain fundamental meaning structures seem to have remained largely the same. This is, in my view, the reason why we can still today recognize the deep emotions of Sapho when we read her poems or understand the ethical dilemmas expressed by Arjuna in the Bhagavad Gita. To genuinely rethink “the relational” requires that the conditions from which all relations are made possible not be understood solely in terms of relationality. Wittgenstein (in Tractatus, 1922) was fundamentally right, in my view, when he claimed that a world without substance would indeed render impossible forming a meaningful picture of that world.

Relational thinking in contemporary sociology and social theory must avoid using worn-out standard arguments against old dualisms as an excuse for merging everything and dissolving existentially pre-given boundaries in the socio-historical world. Using an outdated rationalistic (Cartesian) view of subject/object as a target to reject the specificities of human nature is a theoretical dead-end. For example, Smith (2022, p. 1819) recently wrote: “The ‘posthumanism’ of relationality does not consider human action and meaning as the central focal point nor does it view agency to be the property of either humans or nonhumans.” According to this non-substantialist approach, the existence of everything or everyone is above all a constant “matter of degree,” mediated through a complex web of interdependent agencies, which makes all essential forms, boundaries, and the very meaning of the non-relational totally redundant. This perspective is partly influenced by Latour’s (1994, 2005) actor-network theory and his so-called “flat ontology,” which tends to lead to radical relativization and the dissolution of all boundaries between living beings and non-living artifacts.

In mainstream sociology, individuals have often been portrayed as emerging from the relational dynamics in social fields; Bourdieu and Wacquant (1992, p. 107) made an analogy by using an example from modern physics, claiming that “the individual, like the electron, …is in a sense an emanation of the field.” And a similar view was presented by Harrison White (1992, p. 197), who wrote: “Person should be a construct from the middle of the analysis, not a given boundary condition.” In other parts of the social sciences, such as organization studies, Cooren (2020, p. 16) declared: “Thinking the (organizational) world relationally indeed invites us to think in terms of both/and.” However, my objection would be that we cannot be satisfied with an all-embracive “both/and” only, but have to include a Kierkegaardian notion of the “either/or” in the mix, and put it at the very heart of all mediations—a critical edge characterizing human decisions/indecisions (Roumbanis, 2010). When a person is faced with a difficult decision that will have consequences for the rest of her life, this typically means she cannot choose “both/and,” but must decide between doing “either/or.” Thus, the “either/or” concerns what Scott (2022, p. 208) explained as follows: “making a positive choice toward something demands a negative turn away from something else.” The relational dimension can never exhaust the depths of the single individual; the inner labyrinths of human inwardness will preserve an enclosed space that concerns what Kierkegaard called the “possibility of possibilities” (Kierkegaard, 1844/1980). This is, in my view, an essential aspect of ecstatic relationality. In the next section, my aim is to further explore the notion of mediation by anchoring the multi-dyadic level of relationality in a broader theory of practice and socio-materiality.

A theory of socio-material mediations

Defining the practico-inert field and its existential foundations

Sartre's Critique represents a theory of practice quite different from the sociological contributions of Bourdieu, Giddens, and Latour. In the introduction, I presented a problem definition that pinpoints the deep complexity of mediation, thus a relational ontology that revolves around the movement of individual practices, plurality, the organization of plurality, and the plurality of organizations (Sartre, 1960/2004, p. 80). All human activities take place in what Sartre called the “practico-inert field,” that is, a plurality of concrete socio-historical environments charged with inverted practices and meaning inscribed into matter. This geospatial term contributes to a phenomenological description of the materially arranged places where people interact, embodying the same/different needs, desires, interests, commitments, and opportunities. In one paragraph, Sartre (1960/2004, p. 324) offered a vivid snapshot of the practico-inert field:

I need only glance out the window: I will be able to see cars which are men and drivers who are cars, a policeman who is directing the traffic at the corner of the street and, a little further on, the same traffic being controlled by red and green lights: hundreds of exigencies rise up towards me: pedestrian crossings, notices, and prohibitions; collectives (a branch of the Crédit Lyonnais, a café, a church, blocks of flats, and also a visible seriality: people queueing in front of a shop); and instruments (pavements, a throughfare, a taxi rank, a bus stop etc., proclaiming with their frozen voices how they are to be used). These things,… practical unities made up of man and inert things – these appeals, and these exigencies do not yet concern me directly. Later, I will go down into the street and become their thing, I will buy that collective which is a newspaper, and suddenly the practico-inert ensemble which besieges and designates me will reveal itself on the basis of the total field, that is to say, of the Earth as the Elsewhere of all Elsewheres (or the series of all series of series).

As we can see in this quote, Sartre mixes phenomenological and metaphorical descriptions to illustrate the incredible richness of the practico-inert field, materialized through the innumerable decisions and the hard work of other people, that is, by our contemporaries and by previous generations. All the innumerable configurations of organized and non-organized human interactions in the practico-inert field comprise an “…indefinite multiplicity of third parties” (Sartre, 1960/2004, p. 119).

What, then, is a “third party”? The answer is that everyone and everything can, in principle, be a mediating third party, thus creating a reciprocal relationship in the practico-inert field. Matter is always, in and through humans, the very motor of history; matter shapes societal relationships by providing a vital link that can both connect and separate people (Sartre, 1960/2004, p. 79; cf. Ihde & Malafouris, 2019). Stressing the trinitary nature of all relations is an ingenious way of discovering the latent contradictions and the practical inertia that exist in a great many social situations. Sartre's theory of socio-materiality, however, must be clearly distinguished from non-substantialist approaches (e.g. Cooren, 2020; Latour, 2005; Orlikowski & Scott, 2008). In the Critique, there is no doubt regarding the existential difference between humans and material artifacts. A third party mediates but does not create, an “ontological fusion” that dissolves all boundaries. Most artifacts have their own specific attributes, functions, and forms, which makes them relatively independent and stable in relation to human users, something that led Faulkner and Runde (2012, p. 57) to critically ask the following rhetorical question: “While the needle of a syringe may penetrate the body, does the body penetrate the needle at the same time?”

In the Critique, there is a fundamental qualitative difference between what Sartre identified as “group reciprocities” and “serial reciprocities.” This distinction emphasizes the organizational difference that typically affects people's being-toward each other in different situations, an idea that can be related to Goffman's (1963) sociological analysis of “focused” versus “unfocused” interactions, the latter referring specifically to people's behavior in public places. A group generally implies, among other things, a sense of community, having collective interests, and mutual engagement. Groups entail both consensus and conflicts, and they can vary significantly in size. Sartre's analysis of groups has indeed many similarities with the Simmelian understanding of group dynamics as well as with later sociological studies on the creation, evolution, and dissolution of social groups over time (e.g. Flamino et al., 2021). Other forms of social gatherings in the practico-inert field do not have the same organizational characteristics, for example, crowds or networks, yet they may nevertheless provide the elementary spark that could cause people in these practical ensembles to come together in new “groups-in-fusion” (Sartre, 1960/2004, p. 346). Still, with regard to the organizational elements of groups, there is an essential difference between being a member of a group and being part of an anonymous network. Within an organized group emerges a certain type of reciprocity that non-members cannot possibly share (Ahrne & Brunsson, 2011). However, there is always an immanent and possible conflict between the individual and the common, “which oppose and define each other” as a profound truth, “manifested in new contradictions within the organized group” (Sartre, 1960/2004, p. 583).

Although the main focus in the Critique is on the relational conditions of groups and larger formal organizations in the practico-inert field, Sartre never leaves us in doubt about the importance of “lived experience” (Laing & Cooper, 1964). This is something he explicitly underscores in what we could consider a pre-study to the Critique, namely the book Search for a Method (1957/1968), when he discussed Kierkegaard's and Marx's critical stances against Hegel's dialectical system, a holistic philosophy that swallows up everything in its path, causing all the struggle, irrationalities and paradoxical aspects of human existence to dissolve into objective all-encompassing world-historical mediations. Marx, on the one hand, used his materialistic worldview to turn Hegel's idealistic philosophy on its head. Kierkegaard, on the other, looked everywhere for weapons to aid him in escaping from the dialectical mediations:

He discovers within himself oppositions, indecisions, equivocations which cannot be surpassed: paradoxes, ambiguities, discontinuities, dilemmas, etc. In all these inward conflicts, Hegel would doubtless see only contradictions in formation or in process of development, but this is exactly what Kierkegaard reproaches him for. (Sartre, 1957/1968, p. 11; cf. Roumbanis, 2010)

Being alone together: serial reciprocities and alterity

If belonging to a group offers each individual a certain sense of belonging and reciprocity, then what exactly is “serial reciprocity”? Essentially an ontological category, seriality² is key to understanding the intricate nature of mediation, as it denotes a profound state of alterity between individuals, with socio-materiality very often being the only thing connecting them in certain situations. Seriality, or serial orders, can exist almost everywhere in societies. Money, language, radio broadcasting, and opinion polls are some of the examples Sartre explicitly gave, and all display elements of serial reciprocity. Unlike the reciprocity of a group, the one inherent in serial gatherings is of a very different sort. Individuals who belong to a series may have a common interest, but not as an organized group, only as a plurality of solitudes. A series of people, a “lonely crowd,” is typically defined by mutual separation and anonymity.³

One of Sartre's most famous examples of visible seriality concerns people waiting at a bus stop. He used the queue to illustrate a concrete form of serial order, pointing to its temporary and contingent, yet regulated, manifestation in the practico-inert field. The bus (as part of a public transportation system) unites the people in their common interests as commuters. Seriality highlights a contradiction that exists between the uniqueness of each individual and the interchangeability of each as a commuter. From this point of view, alterity “becomes a negative principle of unity and of determining everyone's fate as other by every Other as Other” (Sartre, 1960/2004, p. 261).

However, this form of serialized gathering at the bus stop can suddenly change into something more meaningful and dynamic—a group-in-fusion—starting, for example, with a particular event that may trigger spontaneous actions or communication. A non-event can also make things happen, like a bus not arriving on time at the bus stop in the early morning, thus generating a common concern among those waiting in the queue. All individuals depend on the bus; they need it to arrive. Absence is an existential modality that can have a decisive impact on human interaction, which illustrates the ecstatic aspect of relationality, or as described by Scott (2022, pp. 197–198) as the lack of a substantive matter, “whether this be physical, material, ideal or symbolic." When a bus is delayed, people might start talking to each other, momentarily breaking the serialized order and giving a small indication of solidarity. When the bus finally arrives, people will almost immediately fall back into the contingent plurality of solitudes (Sartre, 1960/2004, pp. 256–257). Seriality represents different degrees of massification and instrumentalization. Moreover, owing to increasing differentiation and inertia, the serialization of human interaction can appear within organized groups and formal organizations. As Sartre (1960/2004, p. 679) explained, “the group is always marked by the series.” This statement highlights the ecstatic nature of all human relations and entities situated in the practico-inert field. But the question is: How can we transfer these insights to our contemporary societies, in which automation and digitalization have become key elements of organizational practices? Implementation of AI-powered decision-making systems has radically changed the ways in which social groups are created and new serial reciprocities emerge. People are navigating in a practico-inert field where machine-learning algorithms are becoming increasingly sophisticated in analyzing visual data (e.g. facial recognition systems, autonomous military drones), using human language for evaluative purposes (interview chatbots, medical support chatbots), and creating new strategic solutions in contexts marked by strong competition (financial markets, warfare, and sports). This is what some sociologists have called the “society of algorithms,” thus pointing to the mediating aspects of AI in many important decision-making processes that impact the distribution of scarce resources and opportunities (e.g. Burrell & Fourcade, 2022; O’Neil, 2016; Pasquale, 2015; Roumbanis, 2025).

A metaphorological understanding of algorithmic mediations

New algorithmically generated third-party relations

We have now reached the third and final part of the present study. The objective at this point is to theorize algorithmic mediation by focusing on machine-learning algorithms as “third parties” that shape new reciprocities between people in the practico-inert field. Based on my previous discussion of ecstatic relationality and Sartre's theory of mediation, in what follows my aim is to shed light on factors that characterize the relational conditions between humans and AI technologies. By transferring ideas regarding the phenomenologically concrete to the contexts of opaque algorithms, it is important to underscore the manual labor and “the materiality of these seemingly immaterial algorithmic processes as they shape decision-making” (Beer et al., 2024, p. 3).

Here, to focus on relational modalities, I will use a metaphorological approach that illustrates some key aspects of algorithmic mediations. Metaphors can play an important role in general theorizing. In Sartre's Critique, material artifacts are portrayed as having almost magical powers; metaphors like “vampire objects” and “bewitched quantities” are used to describe tension-filled interdependencies that can occur between human practices and the practico-inert field. The metaphor “frozen voices” is used to highlight the communicative and sense-making dimensions entailed in a great many things in people's everyday life (Sartre, 1960/2004, p. 324). I will use these three metaphors—vampire objects, bewitched quantities, and frozen voices—to explore the mediating role of AI systems in organizing individual practices, group configurations, and serial orders. Needless to say, metaphors can lead us astray, and we should keep this in mind when we engage in theory construction and treat metaphors in a careful manner. However, metaphors can have strong heuristic and suggestive powers that help us see phenomena in a different light (Swedberg, 2020). Despite the rather fanciful names of the three metaphors I will use in the following sections, my main intention is to open up an interpretative space for an alternative relational ontology by providing a critical wedge and conceptual viability for understanding the human being-toward different types of algorithmic phenomena.

Metaphor 1: algorithmic systems as “Vampire Objects”

Machine learning models typically need to be trained on reliable and unbiased datasets if they are to provide organizations with augmented decision-making. To work efficiently, they must often undergo significant supervised training, so-called reinforcement learning based on human feedback and fine-tuning (Arkoudas, 2023). A concrete artifact can be interpreted as a “vampire object” in the sense that it requires regular human care and attention, which Sartre described as follows:

To preserve its reality as a dwelling a house must be inhabited, that is to say, looked after, heated, swept, repainted, etc., otherwise it deteriorates. This vampire object constantly absorbs human actions, lives on blood taken from man and finally lives in symbiosis with him. (Sartre, 1960/2004, p. 169)

To generalize from this example, AI systems also absorb human actions; they represent highly complex and opaque forms of inverted practices, designed to perform specific tasks that will determine people's life chances (Beer, 2024; Cevolini & Esposito, 2020; Gerdon et al., 2022; Pasquale, 2015).

Algorithm-based technologies need regular care to function properly; they also need to be monitored by humans to prevent them from becoming biased or erroneous, which could have negative effects (e.g. Gillespie, 2024; O’Neil, 2016; Roumbanis, 2025). But the point is that algorithms absorb human actions in many different ways, for example, during the designing, coding, and training phases. Generative AI systems are known to be “data hungry,” which means they need to be fed with biometric data extracted from real-word contexts if they are to work properly in complex decision-making operations. I argue that algorithmic systems represent a very specific type of vampire object in the way they crave biometric data, social media data, or annotated facial images that are mined and collected as a precious “behavioral surplus” stemming from real human actors (Jacobsen, 2023; Zuboff, 2019). But what exactly distinguishes the vampiric nature of these algorithmic artifacts? Despite the availability of enormous amounts of historical data, the problem today seems to be a lack of enough training data for current machine learning models. Historical datasets are often incomplete, biased or marked by unpredictable risks. In recent years, this challenge has led to the use of so-called “synthetic data” for training algorithms, with the promise of providing an extremely rich source of variability and diversity (Jacobsen, 2024; see also Jordon et al., 2022). But synthetic data are not real, because they are not extracted directly from the empirical world. Instead, synthetic data are generated by algorithms for other algorithms, thus giving rise to quite a peculiar relational dimension. Algorithmically mediated relations add new dimensions to social life, but it is far from clear what the long-term effects will be for the organization of human relations. As Jacobsen (2023, p. 6) wrote, “Synthetic data can, … be understood as a technology that both cuts up the world and binds it together in new ways, generating new actionable correlations, diagnoses, and ultimately futures.” Algorithmic systems may be among some of the most challenging vampire objects humans have invented thus far; the consequences of their deep learning capacities and the influence of synthetic data may eventually have quite odd serializing effects in the practico-inert field. In fact, there is a non-negligible risk that the “data hungry” algorithmic systems can lead to organizations circumventing consent for data use, for instance in facial recognition technologies, because as some scholars have recently pointed out, synthetic data are prone to consolidate power away from “those most impacted by algorithmically-mediated harm” (Deslandes Whitney & Norman, 2024, p. 1). In other words, algorithmic mediations can be things that not only connect and coordinate but also their vampiric effects can also be alienating and harmful in different ways. The agentic properties of algorithmic artifacts can be alienating, an idea that has no real place in Latour's actor-network theory, for example, simply because it dissolves the substantial boundaries between the living and the non-living⁴ to the extreme. To make a deeper sense of alienation, we must instead acknowledge that it can indeed be “present on all levels of experience” (Sartre, 1960/2004, p. 67; see also Axelos, 1976; Jaeggi, 2014). Hence, algorithmic mediations can be fundamentally alienating for various groups of people working with AI technologies because deep learning algorithms constantly absorb human actions and need to be monitored. But algorithmic mediations can also potentially be alienating for the lonely crowed of serialized individuals, all those who do not have the practical abilities or resources to put up resistance against the power of the algorithms they are momentarily exposed to.

Metaphor 2: the “Bewitched Quantities” of algorithmic systems

The metaphor “bewitched quantities” will be used here to theorize the potentially disruptive consequences that floods of algorithmically generated decisions may have for formal organizations as well as the general welfare of ordinary people. Let us take a first look at this metaphor by going back in history. In the Critique, Sartre describes the enormous outflow of Peruvian gold to Spain during the 1500–1650s, demonstrating its non-intended effects on socio-economic relations, thus illustrating the core meaning of “bewitched quantities.”

Every effect of Spanish money was a transformation and reflection of human activity: wherever we find that an action of gold disrupts human relations without being willed by anyone, we discover an underlying profusion of human undertakings, directed towards individual or collective ends and metamorphosed through the mediation of things. (Sartre, 1960/2004, p. 182)

This historical example provides insight into what happens when too much of something leads to undesirable effects that can be extremely difficult for humans to control. Overflows of a particular object like gold can suddenly become a bewitched quantity, without anyone being able to anticipate the event and counteract it before it is already too late. The mediation of things, and the values they entail, can today also be metamorphosed algorithmically, thus creating new kinds of bewitched quantities that may disrupt and even disintegrate human group conditions.

A timely example of algorithmically generated bewitched quantities comes from the world of finance and the development of so-called “high-frequency trading.” Min and Borch (2022) recently discussed how the human traders who used to populate the trading floors have been replaced by deeply interconnected machine learning systems. Today, the AI technologies that are operated by different market actors constitute a highly interactive market ecology. While the algorithms obviously have been developed by human software engineers and are being monitored by human analysts, “the actual decisions to buy or sell securities are made by the algorithms themselves and automatically executed through comprehensive technological infrastructure that […] has also introduced new types of market failures and crashes” (Min & Borch, 2022, p. 278; see also Borch, 2022). The point is that these deeply mediating algorithms can be quite vulnerable, with the failure of one system potentially generating millions of erroneous orders, in turn creating strong organizational turbulences with potentially widespread impacts on societies. Hence, these inverted practices (to use Sartre's terminology)—materialized as AI systems, constructed, designed, and coded by humans—suddenly strike back as bewitched quantities with full power, thus leading to fractured algorithmic mediations. Ultimately, these automated decision-making processes, the bewitched quantities of high-frequency trading, can in the worst case cause an organization to collapse in less than an hour, due to an algorithmic phenomenon that has come to be known as “Flash Crashes” (Min & Borch, 2022, p. 293). Briefly, when some AI systems start behaving erratically for some unknown reason (because of their extreme opacity and speed), and other algorithmic machines respond by withdrawing from trading in an attempt to minimize risks, this may have chaotic effects. These “third-party” relations do not create a seamless process like some organizations scholars have proposed, but cause multiple separations and can have a disastrous impact on societal goods. Although communication is a fundamental part of understanding the agentic properties of algorithmic technologies, today they can be extremely complex, making them incomprehensible and immensely difficult to interpret, even for computer scientists (e.g. Fazi, 2021; Yampolskiy, 2024). The innumerable interactions taking place between autonomous algorithmic systems in high-frequency trading can create an abyss of unpredictability. Thus, if not properly handled, they can lead to the disintegration of groups and serializing of people across contexts.

Metaphor 3: the “Frozen Voices” of AI chatbots

Material artifacts are constantly mediating human relations with their “frozen voices” in different situations, or as Sartre (1960/2004, p. 324) explained, "hundreds of exigencies rise up towards me […] proclaiming with their frozen voices how they are to be used.” This is true, but the interesting thing today is the emergence of algorithmic artifacts that are literarily speaking to us with their frozen voices. Here, I am thinking in particular about AI chatbots. Many public and private organizations are currently using chatbots to decrease waiting times and provide more flexible services to ordinary citizens. However, there are also chatbots that are used for evaluative tasks, to determine who deserves or is entitled to receive a certain opportunity or resource (e.g. social welfare payments, employment, and insurance). Indeed, a growing number of professional judgments are based on recommendations provided by algorithmic systems, for instance, in the area of personnel selection (Bursell & Roumbanis, 2024; Roumbanis, 2025). Job candidates are frequently evaluated and sorted by recruitment chatbots. This socio-technical development has been made possible by the recent breakthroughs in natural language processing (NLP), where artificial neural network computations and hundreds of billions of parameters have provided a “powerful universal language understanding” (Manning, 2022, p. 127). This capacity has transformed crucial aspects of the human being-toward others and radically redefined the meaning of technology as a mediating “third party.” Language is, needless to say, fundamental to most forms of human social interaction, and at present, when chatbots can be trained to use human language in a highly skillful manner to communicate about almost everything, that naturally represents a great leap in AI research (e.g. Arkoudas, 2023).

However, the crux of the matter seems to be whether generative AI systems like ChatGPT or BERT can be said to genuinely “understand” the human meaning they are designed to be processing. This is far from a trivial issue, considering that organizations use chatbots that are based on these systems to evaluate people's answers to several tailor-made questions. As Mitchell and Krakauer (2023, p. 2) remarked, an AI chatbot “could use the word “tickle,” but it has obviously never had the sensation. Understanding a tickle is to map a word to a sensation, not to another word.” But some scholars have argued that understanding can also be interpreted as the ability to relate a word to other words, and to identify pictures that correspond to the meaning of these words, although this necessarily assumes a narrower type of understanding (e.g. Manning, 2022). Humans can indeed talk meaningfully about things they have never experienced in real life, but we would still acknowledge that this is some form of knowledge. Yet we would also have to add that this form of indirect understanding could only be regarded as proper sense-making because of the innumerable layers of lived experience and the human capacity of intuitive-analogical reasoning that can transform this into a deeper understanding. Currently, AIs do not have this capacity.

Hence, the metaphor “frozen voices” captures the substantial difference between language that comes directly from living human beings and the language produced by non-living algorithmic systems. It might be the same natural language, but with different kinds of “third-party” actors using it to mediate organizational processes. To be sure, humans can also use language in a rather cold and disinterested, robot-like, manner, but that will inevitably depend on the organizational context (a strictly rule-following bureaucrat perhaps being an archetypical example). But there is still a significant difference. Although AI chatbots are highly competent language users – able to carry out conversations on arbitrary topics with extraordinary levels of fluency and coherence – some critical scholars have dismissed them as merely being “stochastic parrots” (Bender et al., 2021). Other commentators have claimed, however, that this critique has not been warranted since the entry of GPT-4, while still acknowledging that reasoning and logical soundness remain the real Achilles heels of all currently available conversational AIs (Arkoudas, 2023; see also Cantwell Smith, 2019).

Nonetheless, generative AI has the potential to offer new and creative readings of texts, for example, in the digital humanities, with the aim to surprise and provoke (see, e.g., Evans and Foster, 2024). This is because the outputs of a machine learning model are not necessarily straightforward reproductions of the billions of fragments of human data it was originally trained on. As such, the frozen voices metaphor draws attention also to the intricate interplay between humans and AIs – a form of ‘hybrid intelligence’ (Jarrahi et. al., 2022) – not to be confused, however, with the Latourian actant. Still, there is a crucial difference between a ChatGPT generating a poem and a recruitment chatbot programed to evaluate job candidates.

In any case, what the metaphor “frozen voices” really highlights is the practical impact chatbots have by virtue of their organizational actorhood, thus mediating certain forms of meaning between, for example, expert recruiters and job candidates in the process of personnel selection. Algorithmic mediations do imply classification and separation of people. For instance, the majority of all the hundreds of job candidates applying for a vacant position will probably never meet a human recruiter if they are not making a good impression on the chatbot. And as mentioned above, there are well-known limitations on how much a chatbot can “understand,” which can obviously have direct practical consequences for how it scores the candidates’ performances. For example, a candidate who answers a question by relating a hypothetical scenario may be misinterpreted by the chatbot, thus wrongly assuming the candidate was referring to something he/she had actually experienced in the past. Moreover, with many organizations today using so-called AI video hiring systems, job candidates not only communicate with, but are also actively observed by, the recruitment technologies. And as Drage and Mackereth (2022, p. 89) noted, these AIs utilize speech and facial recognition software; head movements, gestures, micro-expressions, word use, and intonation are all tracked and interrogated to gauge whether a candidate is the right person for the job. This type of AI recruitment system takes the issue of algorithmically mediated evaluations to yet another level.

If we now briefly return to Sartre's distinction between serialities/groups, then what exactly are the implications? Algorithms have indeed become quite powerful “gatekeepers” in shaping group affiliations and organizational routines in the practico-inert field, at least as long as human decisionmakers do not experience “algorithm aversion” and refuse to comply with the AI's recommendations (e.g. Bursell & Roumbanis, 2024; Burton et al., 2020; Kellogg et al., 2020). Nevertheless, chatbots have become key actors in the practice of “constructing people as cases,” which can obscure individual narratives (Kiviat, 2023). Implementation of AI in the recruitment process offers an illustrative example of the serializing impact algorithmic mediations can have in everyday life. Of the many hundreds of job candidates who will encounter a chatbot in an initial screening interview, only a few will pass to the next step and meet a human recruiter. Even fewer will be hired and assigned a new group affiliation. Most job candidates will remain contingent members of an algorithmically generated serial gathering, the “invisible masses” within these digital platforms. It is not necessarily the case that traditional recruitment was better, but the frozen voices of the chatbots have changed the phenomenological experience of communicating and the practical conditions for getting a job. And for all those professionals in charge, the human expert recruiters, we can only add what Stiegler (1998, p. 21) wrote on humanity experiencing the deep opacity of modern technologies: “we do not immediately understand what is being played out in technics, nor what is being profoundly transformed therein, even though we unceasingly have to make decisions.”

Concluding Remarks

In contrast to the prevailing views in mainstream relational sociology, post-phenomenology, and actor-network theory, in the present article I have instead proposed a theory of relationality based on a substantialist philosophy that assumes the relative independency and discreteness of all individual beings. Based on this view, I then described the notion of mediation by emphasizing the ecstatic dimensions of all relations using the philosophical concept of being-toward, and then combined this conceptual framework with Sartre's socio-materialist understanding of mediation. Briefly, this constitutes a relational ontology that I believe provides a more complex and dynamic view of mediations in general. There is certainly a need for a new dialogue regarding the very meaning of relationality and substantialism, and I hope the present article can help move us beyond previous “manifestos” and mainstream critiques against outdated dualisms. We need this discussion today in particular, when generative AIs are transforming organizational practices by replacing humans in crucial parts of evaluation and decision-making processes. The latest inventions in deep learning technology and NLP have developed the agentic powers of artifacts considerably, leading to the industrial production of autonomous vehicles, recruitment chatbots, and AI robotics in surgery, etc. It is perhaps more important than ever before that we, now and then, change our perspective by looking up at the starry sky to feel the same kind of deep admiration and awe Kant once expressed.

The notion of algorithmic mediation is complex, and here I have used a different approach to illustrate the emergence of new socio-technical phenomena. Three metaphors were chosen to shed light on crucial aspects of algorithmic mediations, including both their serializing effects on human relations and their impact on group creations. I used three empirical phenomena to illustrate the metaphors presented: the development of so-called “synthetic data,” the rise of “high-frequency trading,” and the implementation of recruitment chatbots in personnel selection. We do not have to question the idea that many algorithmic systems may be extremely beneficial, but we must also understand how algorithmically mediated processes could generate deeply alienating and disintegrating effects in social life.

Footnotes

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Forskningsrådet om Hälsa, Arbetsliv och Välfärd (Grant number 2022-00625) and Vetenskapsrådet (Grant number 2019-02550).

ORCID iD

Lambros Roumbanis

Notes

Author Biography

Lambros Roumbanis is an associate professor of sociology at Örebro University and a researcher at the Institute for Futures Studies (Stockholm). His main research interests are in the sociology of science and STS, social theory, sociology of algorithms, and contemporary philosophy of technology.

References

Ahrne

Brunsson

(2011). Organization outside organizations: The significance of partial organization. Organization, 18(1), 83–104. https://doi.org/10.1177/1350508410376256

Amoore

(2020). Cloud ethics: Algorithms and the attributes of ourselves and others. Duke University Press.

Arkoudas

(2023). ChatGPT is no Stochastic Parrot. But it also claims that 1 is greater than 1. Philosophy & Technology, 36(3), 1– 29. https://doi.org/10.1007/s13347-023-00619-6

Axelos

(1976). Alienation, praxis, and Technē in the thought of Karl Marx. University of Texas Press.

Beer

(2024). The tensions of algorithmic thinking: Automation, intelligence and the politics of knowing. Bristol University Press.

Beer

Wallace

Burrows

Ciocanel

Cussens

(2024). Valuing the manual: The demarcation of embodied practices within algorithmic decision-making processes. Social & Cultural Geography, 25(10), 1575–1593. https://doi.org/10.1080/14649365.2024.2347872

Bender

E. M.

Gebru

McMillan-Major

Shmitchell

(2021). On the dangers of stochastic parrots: Can language models be too big? In: Proceedings of the 2021 ACM conference on fairness, accountability, and transparency (pp. 610–623). FAccT ‘21. New York: Association for Computing Machinery.

Borch

(2022). Machine learning and social theory: Collective machine behaviour in algorithmic trading. European Journal of Social Theory, 25(4), 503–520. https://doi.org/10.1177/13684310211056010

Bourdieu

(1998). Practical reason: On the theory of action. Stanford University Press.

10.

Bourdieu

Wacquant

(1992). An invitation to reflexive sociology. University of Chicago Press.

11.

Burkitt

(2016). Relational agency: Relational sociology, agency and interaction. European Journal of Social Theory, 19(3), 322–339. https://doi.org/10.1177/1368431015591426

12.

Burrell

Fourcade

(2021). The society of algorithms. Annual Review of Sociology, 47(1), 213–237. https://doi.org/10.1146/annurev-soc-090820-020800

13.

Bursell

Roumbanis

(2024). After the algorithms: A study of meta-algorithmic judgments and diversity in the hiring process at a large multisite company. Big Data & Society, 11(1), 20539517231221758. https://doi.org/10.1177/20539517231221758

14.

Burton

J. W.

Stein

M. K.

Blegind Jensen

(2020). A systematic review of algorithm aversion in augmented decision making. Journal of Behavioral Decision Making, 33(2), 220–239. https://doi.org/10.1002/bdm.2155

15.

Cantwell Smith

(2019). The promise of artificial intelligence: Reckoning and judgment. The MIT Press.

16.

Cevolini

Esposito

(2020). From pool to profile: Social consequences of algorithmic prediction in insurance. Big Data & Society, 7(2), 1–11. https://doi.org/10.1177/2053951720939228

17.

Cooren

(2020). Beyond entanglement: (Socio-)materiality and organization studies. Organization Theory, 1(3), 1–24. https://doi.org/10.1177/2631787720954444

18.

Crossley

(2011). Towards relational sociology. Routledge.

19.

Dépelteau

(2018). The Palgrave handbook of relational sociology. Palgrave Macmillan.

20.

Deslandes Whitney

Norman

(2024). Real risks of fake data: Synthetic data, diversity-washing and consent circumvention. In ACM Conference on fairness, accountability, and transparence. ACM FAccT ‘24, June 3-6 2024, Rio de Janeiro, Brazil. ACM, New York, USA, p. 19.

21.

Donati

Archer

M. S.

(2015). The relational subject. Cambridge University Press.

22.

Drage

Mackereth

(2022). Does AI debias recruitment? Race, gender, and AI’s “eradication of difference”. Philosophy & Technology, 35(89), 1–25. https://doi.org/10.1007/s13347-022-00543-1

23.

Elias

(1978). What is sociology? Hutchinson.

24.

Emirbayer

(1997). Manifesto for a relational sociology. American Journal of Sociology, 103(2), 281–317. https://doi.org/10.1086/231209

25.

Evans

J. A.

Foster

J. G.

(2024). Algorithmic abduction: Robots for alien reading. Critical Inquiry, 50(3), 376–401. https://doi.org/10.1086/728933

26.

Faulkner, P. & Runde, J. (2012). On sociomateriality. In: P. M. Leonardi, B. A. Nardi & J. Kallinikos (Eds.), Materiality and organizing: Social interaction in a technological world (pp. 49–66). Oxford University Press.

27.

Fazi

M. B.

(2021). Beyond human: Deep learning, explainability and representation. Theory, Culture & Society, 38(7-8), 55–77. https://doi.org/10.1177/0263276420966386

28.

Flamino

Szymanski

B. K.

Bahulkar

Chan

Lizardo

(2021). Creation, evolution, and dissolution of social groups. Scientific Reports, 11(1), 17470. https://doi.org/10.1038/s41598-021-96805-7

29.

Gasché

(1999). Of minimal things: Studies on the notion of relation. Stanford University Press.

30.

Gerdon, F., Bach, R. L., Kern, C., & Kreuter, C. (2022). Social impacts of algorithmic decision-making: A research agenda for the social sciences. Big Data & Society, 9(1), 1–13. https://doi.org/10.1177/20539517221089305

31.

Gillespie

(2024). Generative AI and the politics of visibility. Big Data & Society, 11(2). https://doi.org/10.1177/20539517241252131

32.

Goffman

(1963). Behavior in public places: Notes on the social organization of gatherings. The Free Press.

33.

Howard

Borenstein

(2018). The ugly truth about ourselves and our robot creations: The problem of bias and social inequity. Science and Engineering Ethics, 24(5), 1521–1536. https://doi.org/10.1007/s11948-017-9975-2

34.

Husserl

(2001 [1900/1901]). Logical investigations. Routledge.

35.

Ihde

(1990). Technology and the lifeworld: From garden to earth. Indiana University Press.

36.

Ihde

Malafouris

(2019). Homo Faber revisited: Postphenomenology and material engagement theory. Philosophy & Technology, 32(2), 195–214. https://doi.org/10.1007/s13347-018-0321-7

37.

Jacobsen

B. N.

(2023). Machine learning and the politics of synthetic data. Big Data & Society, 10(1), 1–10. https://doi.org/10.1177/20539517221145372

38.

Jacobsen

B. N.

(2024). The logic of the synthetic supplement in algorithmic societies. Theory, Culture & Society, 41(1). https://doi.org/10.1177/02632764231225768

39.

Jaeggi

(2014). Alienation. Columbia University Press.

40.

Jarrahi

M. H.

Lutz

Newlands

(2022). Artificial intelligence, human intelligence and hybrid intelligence based on mutual augmentation. Big Data & Society, 9(2), 20539517221142824. https://doi.org/10.1177/20539517221142824

41.

Jaworski

(2016). Structure and the metaphysics of mind: How hylomorphism solves the mind-body problem. Oxford University Press.

42.

Jordon

Szpruch

Houssiau

Bottarelli

Cherubin

Maple

Cohen

S. N.

Weller

(2022). Synthetic data – What, why and how? The Alan Turing Institute and The Royal Society.

43.

Kahneman

Sibony

Sunstein

C. R.

(2021). Noise: A flaw in human judgment. Little Brown Spark & Co.

44.

Kellogg

K. C.

Valentine

M. A.

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

45.

Kierkegaard

(1844/1980). The concept of anxiety. Princeton University Press.

46.

Kiviat

(2023). The moral affordances of constructing people as cases: How algorithms and the data they depend on obscure narrative and noncomparative justice. Sociological Theory, 41(3), 175–200. https://doi.org/10.1177/07352751231186797

47.

Köchling

Wehner

M. C.

Warkocz

(2023). Can I show my skills? Affective responses to artificial intelligence in the recruitment process. Review of Managerial Science, 17(6), 2109–2138. https://doi.org/10.1007/s11846-021-00514-4

48.

Laing

R. D.

Cooper

D. G.

(1964). Reason and violence. Tavistock.

49.

Latour

(1994). On technical mediation. Common Knowledge, 3(2), 29–64.

50.

Latour

(2005). Reassembling the social. Oxford University Press.

51.

Malafouris

(2013). How things shape the mind: A theory of material engagement. MIT Press.

52.

Manning

C. D.

(2022). Human language understanding & reasoning. Dædalus, 151(2), 127–138. https://doi.org/10.1162/daed_a_01905

53.

Min

B. H.

Borch

(2022). Systemic failures and organizational risk management in algorithmic trading: Normal accidents and high reliability in financial markets. Social Studies of Science, 52(2), 277–302. https://doi.org/10.1177/03063127211048515

54.

Mitchell

Krakauer

D. C.

(2023). The debate over understanding in AI’s large language models. Proceedings of the National Academy of Sciences, 120(13), e2215907120. https://doi.org/10.1073/pnas.2215907120

55.

O’Neil

(2016). Weapons of math destructions: How big data increases inequality and threatens democracy. Penguin.

56.

Orlikowski

Scott

(2008). Sociomateriality: Challenging the separation of technology, work and organization. Academy of Management Annals, 2(1), 433–474. https://doi.org/10.5465/19416520802211644

57.

Pasquale

(2015). The Black Box Society: The secret algorithms that control money and information. Harvard University Press.

58.

Proudhon

P.-J.

(1849). De la Création de L’Order dans L’Humanité ou Principes D’Organisation Politique. Garnier Frères, Libraires.

59.

Rea

(2011). Hylomorphism reconditioned. Philosophical Perspectives, 25(1), 341–358. https://doi.org/10.1111/j.1520-8583.2011.00219.x

60.

Roumbanis

(2010). Kierkegaard och sociologins blinda fläck [PhD. Thesis, in English: Kierkegaard and the blind spot of sociology]. Acta Universitatis Stockholmiensis.

61.

Roumbanis

(2025). On the present-future impact of AI technologies on personnel selection and the exponential increase in meta-algorithmic judgments. Futures, 166(2), 103538. https://doi.org/10.1016/j.futures.2025.103538

62.

Ruin

(2019). Being with the dead: Burial, ancestral politics, and the roots of historical consciousness. Stanford University Press.

63.

Sartre

J.-P.

(1957/1968). Search for a method. Vintage Books.

64.

Sartre

J.-P.

(1960/2004). Critique of dialectical reason: Theory of practical ensembles. Verso.

65.

Scott

(2022). Social nothingness: A phenomenological investigation. European Journal of Social Theory, 25(2), 197–216. https://doi.org/10.1177/1368431020958899

66.

Simmel

(1903/1971). The metropolis and mental life. In: D. Levine (Ed.), On individuality and social forms (pp. 324–340). The University of Chicago Press.

67.

Simmel

(1908/1971). How is society possible? In: D. Levine (Ed.), On individuality and social forms (pp. 6–23). The University of Chicago Press.

68.

Simpson

W. M. R.

(2023). Hylomorphism. Cambridge University Press.

69.

Smith

W. R.

(2022). On relationality and organizationality: Degrees of durability, materiality, and communicatively constituting a fluid social collective. Organization Studies, 43(11), 1815–1837. https://doi.org/10.1177/01708406211035497

70.

Stiegler

(1998). Technics and time 1. Stanford University Press.

71.

Strathern

(2018). Relations. In Stein

(Ed.), The open encyclopedia of anthropology (pp. 1–14). Facsimile of the first edition in The Cambridge Encyclopedia of Anthropology.

72.

Swedberg

(2020). Using metaphors in sociology: Pitfalls and potentials. The American Sociologist, 51(2), 240–257. https://doi.org/10.1007/s12108-020-09443-3

73.

Verbeek

P.-P.

(2012). Expanding mediation theory. Foundations of Science, 17(4), 391–395.

74.

White

(1992). Identity and control: A structural theory of social action. Princeton University Press.

75.

Wittgenstein

(1922). Tractatus logico-philosophicus. Kegan Paul.

76.

Yampolskiy

R. V.

(2024). On monitorability of AI. AI and Ethics, 1–19. https://doi.org/10.1007/s43681-024-00420-x

77.

Zuboff

(2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. Profile Books.

On algorithmic mediations

Abstract

Keywords

Introduction

Rethinking the notion of mediation

Substance and ecstatic relationality: being-toward what, how, and why?

A theory of socio-material mediations

Defining the practico-inert field and its existential foundations

Being alone together: serial reciprocities and alterity

A metaphorological understanding of algorithmic mediations

New algorithmically generated third-party relations

Metaphor 1: algorithmic systems as “Vampire Objects”

Metaphor 2: the “Bewitched Quantities” of algorithmic systems

Metaphor 3: the “Frozen Voices” of AI chatbots

Concluding Remarks

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

Notes

Author Biography

References