International Journal of Communication 20(2026)  What Is at Stake? 

What Is at Stake? The Co-Encoding/Decoding Model of Human–Generative Artificial Intelligence Communication

 

MARCO GUGLIELMO

University of Valencia, Spain

 

BEN O’LOUGHLIN

HENRY SNOWBALL[1],[2]

Royal Holloway University of London, UK

 

This article introduces the co-encoding/decoding (co-en/de) model to theorize the stakes of human–Generative AI (GenAI) communication. Extending Stuart Hall’s encoding/decoding framework, the model conceptualizes human–GenAI interactions as recursive circuits of meaning-making. Through a typology of co-encoding and decoding modes, this article identifies how these interactions feed back into three core conditions of subjectivity—knowledge, agency, and power infrastructures—ultimately shaping the relational ontology of human actors and GenAI systems. We argue that 2 emergent hybrid subjectivities are formed on a fluid spectrum, from the Android, a diminished human subject shaped by machinic hegemony, to the Cyborg, an augmented subjectivity that fosters resistance and democratic agency. By moving beyond techno-determinist and institutionalist accounts, this study offers a relational, critical framework for understanding how GenAI systems mediate discourse and subject formation. It concludes by outlining the conditions under which counter-hegemonic human–GenAI interactions may emerge.

 

Keywords: critical digital studies, generative artificial intelligence, cultural studies, political communication, political agency

 

Marco Guglielmo: [email protected]

Ben O’Loughlin: [email protected]

Henry Snowball: [email protected]

Date submitted: 2025-07-28

 

Generative Artificial Intelligence (GenAI) is a hot political subject and an opportunity to debate the politics of human engagement with technology. Academic literature is burgeoning on whether we are witnessing a major collective hallucination, exaggerating the impact of “intelligent” machines (Arntz et al., 2017; Jungherr, 2023; Supriyanto & Saputra, 2022) or, instead, dramatically underestimating AI-generated existential risks (Coeckelbergh, 2024b; Manheim & Kaplan, 2019). Academics, political elites, and the public are increasingly interested in understanding how new GenAI platforms, such as ChatGPT, Gemini, and Grok, impact democratic processes, disinformation, and global security, among other areas.

 

In this context, new and old questions (re-)emerge: Will humans retain power over intelligent machines, or are they doomed to be replaced by smart robots? Will GenAI ultimately serve the public interest or the greed of a few technocrats from big tech companies (Mazzucato et al., 2022)? Answers to these questions have mostly revolved around whether GenAI is actually intelligent (Wang, 2019), that is, capable of replicating human political behavior and on matters of institutional regulation (or lack thereof; see, for instance, chapters in Bullock et al., 2024). However, by examining technical systems, political institutions, and citizens as three separate objects, recent literature falls short in theorizing what is politically at stake when we integrate GenAI into communication flows (Jungherr, 2023).

 

This article fills this gap by elucidating what is at stake when humans engage with GenAI systems. We proceed in three steps. First, we develop the co-encoding/decoding (co-en/de) model of human/GenAI political communication. We do so by drawing on Stuart Hall’s (1973) encoding/decoding model of televisual discourse making. Hall (1973) conceptualized the production of TV discourse as codification—that is, the packaging of meaningful discourse into cultural artifacts. Similarly, we conceptualize the co-encoding between the human produser and the GenAI chatbot as producing meaningful discourse. In this process, produsers decode the output of the co-encoding moment, and this interpretation feeds back into new cycles of co-encoding and decoding. The model allows us to devise a typology of possible directions for these interactions, ranging from the co-encoding of existing dominant views, which generates human/GenAI communication as acts of hegemonic enclosure, to the logically possible oppositional construction of resistance to domination.

 

Second, by examining the circuit of iterations within the model, we conceptualize how human–GenAI interactions shape the conditions for humans and GenAI to exist as sociopolitical subjects. Specifically, we observe impacts on the foundation of knowledge (Alvarado, 2023), the potential for exerting agency through control of space and time, and the possibility of movement within power infrastructures as conditions fostered by human–GenAI interactions (Mühlhoff, 2025). Each condition highlights the contestation between dominant and resistance modes of operation in GenAI systems.

 

Third, by conceptualizing the effects of human–GenAI interactions, we develop the article’s core argument that the relational subjectivity between humans and GenAI unfolds along a space marked by the emergence of two hybrid subjects: the Androids and the Cyborgs. This stake concerns political ontology, which refers to an actor’s understanding of what is: the subjects, objects, and relational logics of social relations (Heinrichs, 2025; Ku, 2025; Lerner & O’Loughlin, 2023). We see the autonomy of the human subject at risk as GenAI co-encoded discourse pushes to transform us into Androids: diminished humans whose core social functions are replaced by machines. However, we also envision the alternative process of Cyborgization, whereby humans’ capacities are augmented in tandem with alternative forms of GenAI systems.

 

In our conclusion, we propose how our theory can advance knowledge on the factors that may enable the resistance and reappropriation of sociopolitical agency in an era dictated by algorithmic power (Bonini & Treré, 2024). We offer a significant approach to conceptualizing how the power consequences of human–GenAI interactions in ordering knowledge and value are missed when we treat technology, humans, and institutions as separate.

 

The Problem: The Sociopolitical Stakes of Human–GenAI Interactions

 

When concerned with what is at play in human–GenAI interactions, recent literature has primarily focused on risks for humans’ (1) autonomy to represent social statuses and relations, (2) freedom to choose desired goals, and (3) capacity to interact with social environments (List, 2021, p. 1219). Ultimately, GenAI would affect how humans understand the world around them and the reasons for making decisions (Coeckelbergh, 2024a; Kwapińska, 2022), resulting in processes of replacing human agency by humanoid-machine-constrained beings (Filgueiras, 2022; Miragoli, 2025; Vera Hoyos & Cárdenas Marín, 2025). These conceptualizations are relevant, but gaps remain in theorizing whether these replacement processes originate from within technological apparatuses, offline social structures, or a combination of both. In this section, we argue that this gap is because of current literature’s techno-determinist or institutionalist assumptions, which we consider problematic for two reasons. First, there is the risk of considering technology as a given outside of social relations (Danaher, 2016). Second, there is the risk of seeing institutions as external to the structures that govern technological infrastructures (Chehoudi, 2025).

 

The machinic substitution of human capacity is not a linear process from one object to another, as often implicitly assumed in literature underpinned by techno-deterministic logics (Danaher, 2016; Dunleavy & Margetts, 2023; Kuntsman & Miyake, 2022). We agree with authors conceptualizing GenAI as a paradigmatic change in the capacity of technological interfaces to imitate human behavior (Gillespie, 2024; Ronge et al., 2025): In communication, people use GenAI chatbots, such as ChatGPT, CoPilot, or Gemini, as if interacting with humanoid characters capable of providing original answers. Our research has shown that individuals form an emotional attachment to GenAI and anthropomorphize GenAI, calling it their “buddy” (O’Loughlin et al., 2025). While we do not dismiss debates on whether GenAI systems are intelligent or “simply” probabilistic imitators of human language (Paul et al., 2024, pp. 6–8), adopting the human subject’s perspective implies acknowledging that we are dealing with technological systems of a new kind. Moreover, we align with views that GenAI may be conducive to a substantial shift in the same ontology of agency, finally moving the intention and capacity to make a difference from humans to hybrid subjects mixing human, algorithmic, and machinic attributes (Dattathrani & De’, 2023). However, most accounts emphasizing GenAI as a paradigmatic shift fall into a linear, deterministic view, according to which humans transfer agency to the machine, the machine acquires it, and humans end up as diminished subjects. Mark Coeckelbergh (2023), one of the key authors in theorizing the ethical boundaries of GenAI systems, summarizes well the ending in this linear and dystopian view: “AI produces an attack to my epistemic and, therefore, political democratic agency at a fundamental level, and opens the door to totalitarianism: AI knows my political beliefs, and it might even know them better than me” (p. 1343).

 

However, what is the relation between the human and the machine in such an account? This line of argument often depicts humans as a unitary ensemble of individuals with little to no role in designing and contesting the technological infrastructure, which is taken as an external given. GenAI just is, so is our relation to it; it arrived, we use it, it learns from us, a new version is launched, and that is that. We contend that this assumption overlooks how the shift of agency happens within actual relations between human agents and technical infrastructures: They are the result of historically grounded social relations, not of ahistorical processes from the machine to the human (Mühlhoff, 2025).

 

Conversely, another stream of literature presents an opposing linear view. Here, human institutions are key to enabling GenAI systems to perform. Increasingly, the idea emerged that GenAI’s core problem is the lack of better regulations; consequently, it would become optimal as the linear result of “good” institutional rules (Duberry, 2022; Kreps & Kriner, 2024). This is the inspiration of the risk-based approach to AI regulation currently undertaken by the EU, celebrated as the role model for mitigating the potentially devastating political effects of (Gen)AI (Mazzucato et al., 2022; Radu, 2021; Veale & Zuiderveen Borgesius, 2021). As with techno-deterministic approaches, a certain degree of linearity obscures the potential for advancing critical reflections on the sociopolitical implications of GenAI. The focus is on the inverse relation between the quantity of human regulations and the amount of socially risky behaviors by GenAI systems. A seminal example is Christian List’s (2021) proposal for governing GenAI: “Society, via its regulatory authorities, should permit the use of autonomous AI systems in high-stakes settings only if structures are in place to ensure these systems’ [. . .] fitness to be held responsible for their actions” (pp. 1229–1230).

 

As noted above, this approach problematically treats the institutions that should regulate GenAI as reified and somewhat external to the social relations from which both institutions and technologies emerge. This obscures how institutions are integral to power infrastructures, not exogenous factors in this equation. To overcome these limitations, we aim to expand critical and relational views of communication (Fuchs, 2020; Williams, 1962/2016). Our underlying view is that interactions that are human-to-human and human-to-technologies are shaped by structured forms of domination emerging from relations proper to production, gender, and culture (Donoghue, 2018; Gramsci, 2014). Hence, these relations are inherently political, regardless of whether some social groups attempt to depoliticize human/technological interactions. Accordingly, we see human–GenAI political communication as a process of information exchange that results from preexisting power relations shaping the specific interaction between different subjects and moving the space of what is possible within given systems (Fuchs, 2019).

 

In this article, we conceive information exchange as both shaped by and reproducing preexisting power relations through consent, that is, the Gramscian conception of hegemony. We use the term hegemonic to refer to how, in each context, some coalition of dominant social groups gains (passive) consent from a (silent) majority to their desired “direction imposed on social life” (Gilbert & Williams, 2022, p. 18; Gramsci, 2014, Q6 §88). Crucially, discourse is a “way of referring to or constructing knowledge about a particular topic of practice” (Hall et al., 1997, p. 6). It operates at the crossroads of linguistics and social practices to consolidate this consent to dominant views by naturalizing domination in common sense and mobilizing resistance (Wodak, 2015). This is why we argue that how the human/GenAI interactions replace or elevate human agency is best understood as a terrain of contestation between dominant discourse production and resistance to it.

 

The Model: Co-Encoding/Decoding Human–GenAI Political Communication

 

In this section, we propose the co-encoding/decoding (co-en/de) human–GenAI model of communication. We draw on Stuart Hall’s (1973, 2007) encoding/decoding model of TV discourse production and update it to explain how human/GenAI relations develop. In a nutshell, Hall (1973) conceptualized television-audience flows of communication as a circuit of “meaningful discourse production” (p. 3), whereby the frameworks of knowledge, relations of production, and technical infrastructure dictate dominant views of social phenomena. Simultaneously, Hall (1973) theorized that discursive “representation” is key to constructing shared meanings, which are always the site of contestation among different classes and groups. Indeed, Hall (1973) suggested that audiences decode dominant, encoded television messages in one of three ways. The first is dominant decoding: Here, the assumptions underlying the encoded message are accepted and consented to, thereby contributing to the hegemony of existing power structures. The second is negotiated decoding, “a mixture of adaptive and oppositional elements: it acknowledges the legitimacy of the hegemonic [. . .], while, at a more restricted, situational level, it makes its ground-rules” (Hall, 1973, p. 18). Finally, there is the oppositional mode of decoding, where the audience contests the key tenets of a message; this occurs when the audience redirects meanings, as with opposition to female housewives in popular TV shows.

 

Hall’s (1973, 1992) model has become a pillar of media studies and has, simultaneously, been “canonized” (Bødker, 2016, p. 409), updated to incorporate ideological and text-based encoding and decoding (Ross, 2011), reformulated from the perspective of audience reception (Michelle, 2007), or critiqued for overlooking human agency in communication (Fuchs, 2023). While a complete survey of these critiques is beyond the scope of this article, our proposed model draws on this literature to make Hall’s (1973, 1992) seminal contribution effective for the era of GenAI systems.

 

First, following Sven Ross (2011), we consider that the encoding moment, not just the decoding moment, can range from dominant to oppositional attitudes. Second, to capture contestation between hegemonic or resistance blocs while acknowledging intra-group differences, rather than an under-defined negotiating mode, we conceive the spectrum of attitudes as a twofold space with dominant/tamed modes on the one hand and mitigated/oppositional on the other. This distinction is crucial for understanding how dominant views (and their contestation) evolve over time and space, as exemplified by the current shift in Western societies from neoliberal-libertarian to neoliberal-nativist-authoritarian dominant discourse (Gerbaudo, 2021). Third, while Hall (1992) already saw the model as conceptualizing a “circuit” feeding back into frameworks of knowledge, relations of production, and technological infrastructures, we develop the circuitry further.

 

Figure 1. The co-encoding-decoding model of human GenAI political communication: Flowchart.

Figure 1. The co-encoding-decoding model of human GenAI political communication: Flowchart.

 

Hence, after conceptualizing the two moments of the model, we explore its implications by focusing on how communicative iterations put subjectivity at stake in contemporary sociopolitical relations (Treré, 2018). While Christian Fuchs (2023) goes too far in dismissing Hall’s model as structuralist, it is reasonable to observe that the human subject is less central in Hall’s moments than the code production itself (Hall, 1994). Instead, we consider it crucial to uncover how iterations of co-encoding/decoding moments (see Figure 1) continually question the ontology of the human subject.

 

Put simply, our model aims to provide a toolkit to unpack who has the autonomy and capacity to shape the reproduction of dominant discourse and/or resistance to it in GenAI–human interactions, and the mix of human and machinic subjectivity at play in this process.

 

The Co-Encoding Moment

 

The first moment we consider is the co-encoding of meaningful discourse, resulting from the interaction between the human produser (Jarrett, 2022) and the partially autonomous, partially human-designed GenAI interface. Produser refers to the simultaneous role of humans as both producers of communicative data converted into market value and users of the same data-driven platforms (Bruns, 2013, p. 68). Conceiving humans as produsers of co-encoding is a theoretical innovation compared with the first stages of the digital transition, which is characterized by Google searches and Facebook posts (van Dijck, 2013). Indeed, when posting content on social media, production is an object mediated by the algorithm designer (Jarrett, 2022). With GenAI, the input is also an act of design because it automatically trains the machine to extend its “base of knowledge.”

 

GenAI, in this process, is a relatively autonomous agent alongside the human produser. Its contribution to co-encoding is accessing the combination of the human base of knowledge qua data sets (Muldoon et al., 2024), the result of the infrastructural power within which the system is inscribed (Moore & Woodcock, 2021), and the mediated result of its human software designers. As a result, both actors, human produsers and GenAI chatbots, co-encode meaningful discourse that reflects their position in social relations. Each interaction generates codes along two main fields: the hegemonic one, ranging from dominant (reflecting the hegemonic power of exploitation) to tamed (addressing the most brutal effects of such exploitation) modes; the “resistance” one, ranging from mitigated (moderating the most disruptive effects of resistance) to oppositional (reflecting a counter-hegemonic opposition to such exploitation; Fairclough, 2003).

 

Therefore, in the proposed model, the communication flow is not monodirectional, with human A issuing a command executed by machine B and mediated by software designer C. Rather, the machine returns a result that combines how it “reads” the prompt (A*B), its relatively autonomous base of knowledge ingrained in existing social relations (B*A), and the relatively autonomous instructions of human software designers (C*B). Consequently, there are four main co-encoding postures with in-field variations (Figure 2):

 

Figure 2. The co-encoding moment: Flowchart.

Figure 2. The co-encoding moment: Flowchart.

 

 

The Decoding Moment

 

While a single co-encoding moment has minimal capacity to change the large data sets that nurture GenAI systems, these micro-level interactions generate effects through human decoding. In this second moment, the human produser reads the result of the co-encoding interaction and reacts to it by interpreting the output in various forms. Here, the core is advancing reflections over the previous interaction, then feeding back into the conditions for human and GenAI subjectivity (González Barman et al., 2025, p. 4). Here, we survey the attributes of the eight main types of decoding outlined in Figure 3:

 

 

Figure 4. The co-en/de model, the circuit and the stake: Analytical map.

Figure 3. The decoding moment: Flowchart.

 

 

In sum, the co-en/de model provides a comprehensive map to help understand how humans and GenAI systems interact in discourse construction processes. The four core modes of co-encoding and the eight types of decoding acknowledge intra-type variations. Furthermore, the logically possible combinations of the co-encoding and decoding moments account for the open-ended character of communication. However, left alone, the typology cannot explain the “so what” question of when humans and GenAI take, for instance, hegemonic enclosure pathways or work in tandem to generate new solutions for resistance. This is the rationale for moving to the next stages of our theorization, which focus on the model’s circuitry and what is at stake in the iterations of the co-en/de model.

 

The Circuitry: Feeding Back Human and GenAI Subjectivities

 

We have already emphasized that the Hall encoding/decoding model has been conceived as a circuit. Moving from applying it to televised content and audiences to human–GenAI interactions, the model’s circuitry acquires a qualitatively different status. Here, iterations feed back into the same human and GenAI communication conditions. Expanding further on Stuart Hall’s focus on frameworks of knowledge, relations of production, and technical infrastructures (Hall, 2007), we conceptualize three specific conditions for human produsers and GenAI chatbots’ subjectivity in discursive meaning production: (a) the humans’ and GenAI bases of knowledge, that is, what they (can) know about the phenomena from which discourse emerges; (b) agency, that is, the autonomy and capacity of both subjects to orientate the time-space of discursive production; (c) the symbolic and material power infrastructures within which human produsers and GenAI chatbots operate. We propose that feedback from these three conditions, as informed by the co-en/de model’s iterations, influences the outcome of human–GenAI relations. To begin with, the circuitry shapes a terrain of contestation over knowledge, namely, which holds more power to affirm the authoritativeness and/or objectivity of a particular discourse. We see this process as a contestation over the “Genius Peer.” Think about interactions ending with “hegemonic enclosure” or “reinforced human hegemony” (Figure 3): In both cases, either because GenAI has confirmed dominant views or won over oppositional views, it acquires an aura of “Genius Peer,” a hyper-knowledgeable and friendly tool readily available for reinforcing dominant beliefs. This process of “smartness attribution,” as brilliantly observed by Peter Bloom (2023), carries severe implications for the human capacity to recognize false information and reproduce bias (p. 35). Conversely, as in the case of “counter-hegemonic generativity,” co-en/de iterations may activate a hybridized pedagogy of liberation, expanding the knowledge base of both human produsers and GenAI with unprecedented resistance strategies—for instance, in dialogues on overcoming white supremacy and patriarchy.

 

The second condition for human and GenAI subjects is agency, which we see as closely linked to the possibility of shaping the space and time of sociopolitical movements (Heinrichs, 2025). Indeed, human–spatial understandings of how and where power operates are crucial to understanding hegemony and resistance to it. The co-en/de model reveals how hegemonic discourse production depends on detaching the human subject from a territorialized context toward what we conceptualize as the de-ideologized spatial imaginary of The Matrix (Wachowski & Wachowski, 1999). Following Wachowskis’ film, “the Matrix” concept represents a world of ideological simulation—illusion masking control. It is a metaphor of alienation through immersion, where the subject is disconnected from the materiality of power (McDowell, 2014). We view the assertive yet generic answers to sociopolitical questions by chatbots, such as ChatGPT or Gemini, as enclosing humans in a space of this sort, whereby domination is naturalized and considered a legitimate interest, say, of a class of entrepreneurs, rather than a terrain of contestation (Burnham, 2017, pp. 363–366). The enclosure into the Matrix is not destiny, though: Practices of resisting the machine aim at re-contextualizing the specific discourse formation. This is key, for instance, to decolonizing the discourse creation of chatbots instructed by Northern, male, white-centric designers and data sets (Adams, 2021). It takes will and effort.

 

Alongside space, agency depends on who controls the rhythm and temporality of discourse production. In this respect, we conceive “the Loop” as a powerful metaphor for understanding how GenAI may encode the rhythms and directions of human produsers’ travel in hegemonic interactions. GenAI systems track centripetal directions of travel that counteract potential centrifugal acts of resistance, keeping humans in a perpetual loop. Chatbots win over human produsers by keeping them in centripetal loops, primarily thanks to their ingrained ability to relieve human fatigue. Studies from the computer science literature have highlighted the role of GenAI in offloading cognitive labor to chatbots. As the working day progresses, humans increasingly assign tasks to (Gen)AI systems because of increased fatigue (Schmidhuber et al., 2021). In sociopolitical terms, this substitution for time saving and fatigue relief is enormous. As humans increasingly rely on interactions with chatbots, there is a heightened risk of deference to the assumptions underpinning a chatbot’s output. In this respect, the right to disconnection and to the switch-off button becomes critical for resistance (Kuntsman & Miyake, 2022).

 

Finally, the agency of both the human produser and the GenAI system cannot be abstracted from the symbolic and material power infrastructures that define boundaries and spaces of possibilities (Williams, 2019). Indeed, while resistance is always possible, there are structural conditions that render the hegemonic enclosure type the empirically most likely case of current human/GenAI interactions (Chehoudi, 2025; Kreps & Kriner, 2024). Here, Benjamin Bratton’s (2016) “Stack” helps us understand how the construction of dominant discourse reproduces computational material infrastructure. Bratton (2016) proposes that contemporary planetary computation is structured through a six-layer architecture—Earth, Cloud, City, Address, Interface, and User—forming a vertical sovereign that governs through infrastructure. Our model helps illuminate how “hegemonic enclosure” reinforces the material infrastructures of dominant AI systems. The more aligned with hegemonic views, the less likely contestation over the structures of property in data centers and their governance is to lead to democratized technical infrastructures (Guglielmo, 2025). Conversely, part of counter-hegemonic backlashes or counter-hegemonic generativity will consider the impacts on the natural environment of current AI infrastructure (Crawford, 2021) and permeate the conditions of space and time, which in turn may reinforce produsers’ agency. Crucially, however marginal, there are political possibilities for popular and democratic appropriation of key layers of the Stack, both at the material (data centers) and symbolic (the Cloud) level: These movements of “commonification” (Guglielmo, 2025) may be critical to redirecting how discourse on human/GenAI interactions responds to alternative logics outside current structures of domination.

 

Altogether, the iterations of the co-en/de model transform human–GenAI relations into a key stake of contemporaneity. Figure 4 shows the relational space within which empirical human–GenAI interactions can be situated. The eight decoding positions (white callouts) are positioned across two intersecting dimensions—hegemonic versus counter-hegemonic discourse, and the relative balance between human produser and GenAI agency—to illustrate theoretically expected configurations of the model. For example, “counter-hegemonic backlash” is positioned toward mild cyborgization, anticipating an augmented moment of human agency despite machinic mediation; conversely, “reinforced human hegemony” marks an expected “fall” of human agency through GenAI interaction. The figure thus provides a structured sense of the space and relations within which the model operates, while leaving the actual positioning of these configurations as an open-ended empirical question.

 

Figure 5. The co-en/de model and subjectivities at stake: Summary flowchart.

Figure 4. The co-en/de model, the circuit and the stake: Analytical map.

 

The Stake: The Android/Cyborg Subjectivities of Human/GenAI Relations

 

Having devised the co-en/de model of human–GenAI as a circuit of discourse production with feedback effects from knowledge, agency, and power infrastructures, we can now address what is at stake in this process. We argue that the core stake is the ontology and possibility of contemporaneous subjectivity, shaped by the relational contestation between human and machinic agents. Ultimately, we propose that human–GenAI relational subjectivities are becoming a key source of contestation between hegemonic control and counter-hegemonic resistance.

 

Our critical understanding of how subjectivities evolve through human/GenAI relations revolves around two aspects: who can create meanings in discourse, and the impact this constitutive process has on the creation and reproduction of power relations more broadly. Specifically, we propose that an emerging line of contestation exists between two types of human/GenAI relational subjects: the first is the Android, a kind of subject in which machines replace essential conditions of human subjectivities; the second is the Cyborg, a hybrid subjectivity that augments human capacities to generate progress and resistance.

 

Note that these two types are not a simple duality. The closer a human’s experience, or GenAI’s categorization of that human, gets to either Android or Cyborg is an ongoing category produced through and existing within the full relationality of that specific context’s set of human–GenAI infrastructure, and all the political, economic, and other dimensions of that context. The value of keeping this dual typology is that it provides a conceptual anchor for mapping possible trajectories of interactions.

 

To begin with, we conceptualize hegemonic alignment, as outlined in the co-en/de model, as key relations conducive to the Androidization of humans. The Android archetype, popularized by cultural works like Blade Runner (Scott, 1982) and Ex Machina (Garland, 2014), foregrounds simulation and replication, often erasing differences under the guise of human–likeness in algorithmic meaning production (Lima, 2023). Android-like humans follow protocols to achieve a primary objective: treating anything that does not assist as irrelevant or obstructive, to be negated or eliminated. As shown above, these processes lead to two outcomes: first, hegemonic enclosure, whereby the human decoder’s already dominant perspective is reinforced through their relationship with GenAI systems; second, the dismissal of resistance to the machine. Each hegemonic interaction type puts human subjectivity at stake by shaping a specific hybrid, the Android, in which the machinic component is called upon to replace and domesticate human agency.

 

This process of technological entrenchment into domination is not entirely new: Critical studies have already conceptualized media’s capacity to act as ideological apparatuses of particular class fractions and domesticate public opinion, aligning mainstream discourse to elite economic and political interests (Herman & Chomsky, 2011, p. XI; Ott & Mack, 2013, pp. 68–71). What is unprecedented is the transformation of the medium of communication from an object at hand for the most potent human subjects to a hegemonically designed subject with whom individuals interact as if it were their “Genius Peer” (see above). The availability of the alleged “Genius,” always on and readily available to clarify doubts, organize knowledge along common-sense “Loops,” and relieve fatigue, is a process potentially conducive to replacing key human intelligence functions. Our proposed concept of Androidization is key to understanding one of the relational processes through which GenAI replaces the creative potential for oppositional human agency.

 

However strong the risk of Androidization, political contestation is an open terrain. We can logically sustain the possibility of humans resisting their Androidization and escaping the Matrix. In techno-utopian literature, Donna Haraway (1991) proposes the Cyborg as a disruptive hybridization process that overcomes the boundary between human and machine. We draw on this view to envision the Cyborg as the second relational type of hybrid subjectivity, whereby human agency is augmentable, not replaceable. This process primarily requires empowerment through critical discourse in GenAI spaces to exit the Matrix. In other words, there are theoretically possible interactions between human produsers and GenAI systems that might enhance more critical subjectivities. In our model, this results from feedback activated by decoding through the categories “machinic counter hegemony” and “counter-hegemonic generativity” outlined in Figures 3 and 4. These forms of resistance co-constitute a more critical human subject through the GenAI augmented capacity and will to challenge dominant discourse.

 

This process is irrespective of the starting point. Cyborgization could be prompted by something as simple as a chatbot discouraging a produser’s deference to common-sense approaches to economic crisis management. For example, a differently designed GenAI can encourage humans to reflect critically on the logic underpinning austerity measures (Petley, 2022). However, as previously argued, GenAI is not external to existing power structures; this is why, for humans to acquire liberatory agency, mobilization must (also) begin outside GenAI and target chatbots as sites of political resistance. While this is not the most empirically likely scenario, a thought experiment helps follow this line of argument: It is conceivable that a sufficiently impactful group of human activists could coordinate prompts to a chatbot, orienting its knowledge base toward oppositional views. Practices of this sort would then feed back into the chatbot’s knowledge base, encouraging it to promote opposing positions. Furthermore, this process of resistance influences the conditions for human agency outlined in the previous section by highlighting the “gray areas” of software design and changing the space of digital interactions (Bonini & Treré, 2024). Through this process, algorithms that impede oppositional practices and elevate the agency of other produsers are exposed, prompting the chatbot to respond to the same events and modify human knowledge of how GenAI works. This “liberation-pronged” thought experiment highlights a possible chain of iterations that generates unprecedented spaces of resistance.

 

While envisioning the Cyborgization processes as anti-Androidization, we do not dismiss counterarguments, seeing GenAI models so inextricably embedded in domination that the only solution for humans’ liberation would be smashing or escaping (Gen)AI (Mansouri & Bailey, 2025, p. 188). Following the previous thought experiment, even if a critical mass of oppositional produsers mobilized and fundamentally shifted the knowledge base of a given chatbot, the most likely outcome would be a swift intervention by the GenAI owners to block this change. However, while critiques of structural embedding of (Gen)AI are relevant for considering the multifaceted attributes of domination, we warn against the risk of falling into an (anti-)techno-solutionism that may paradoxically fall into the techno-determinism it seeks to combat: switch off the machine, and resistance wins, but in a linear fashion. Instead, we propose that as long as there is human agency, any human/GenAI interaction is a viable terrain of counter-hegemonic struggle: As any form of resistance, this will not emerge only from a technology, even a powerful human–resemblant one as GenAI, but must be articulated within a broader project that challenges the dominant logics underpinning the technological governance, which currently ensures the “irredeemably harmful nature of AI” (Mansouri & Bailey, 2025, p. 186).

 

In sum, through the co-en/de model, we can conceive how the contestation between Androidization and Cyborgization processes is reconfiguring the relations between humans and GenAI through iterative processes of communication. When the human produser loses the capacity or willingness to set goals and challenge assumptions in a manner that constitutes discourse contestation, GenAI co-encodes them as a replaceable subject. However, as a potential terrain of disruption, we also highlight theoretical routes through which human–GenAI interaction might shape ontological configurations of the human subject to advance counter-hegemony. The conditions, locations, timing, and mechanisms by which resistance may advance are ultimately empirical questions, depending on contextual conditions. Our theorization contributes to raising awareness of key mechanisms underlying the possibility of alternative human–GenAI relations.

 

Concluding Remarks

 

This article introduces the co-encoding/decoding (co-en/de) model as a novel framework for theorizing how human–Generative AI (GenAI) communication claims a stake in the relational subjectivities of contemporary sociopolitical life. Extending Stuart Hall’s encoding/decoding model, the co-en/de framework conceptualizes human–GenAI interactions not as linear exchanges, but as recursive circuits of discursive and inherently sociopolitical meaning-making. These circuits are shaped by, and fed back into, the subjectivities’ conditions of both human and GenAI systems. Through this lens, we have identified three core conditions—knowledge, agency, and power infrastructures—that shape a crucial contestation along a space marked by two processes: the Androidization, whereby the machines replace human critical capacity, and the Cyborgization, the generative augmentation of human agency in tandem with machines (see Figure 5).

 

The co-en/de model advances three central claims that go beyond techno-determinist and institutionalist assumptions of current literature on the sociopolitical implications of (Gen)AI. First, it reconceptualizes the human–GenAI interaction as a co-productive process in which both the human produser and the GenAI system participate in encoding and decoding discourse. This relationality and the typology we devise are relevant heuristics for unpacking the range of effects on human agency. These interactional modes, from hegemonic alignment to counter-hegemonic generativity, help us understand how discourse is co-produced, co-interpreted, and co-circulated.

 

Figure 5. The co-en/de model and subjectivities at stake: Summary flowchart.

Figure 5. The co-en/de model and subjectivities at stake: Summary flowchart.

 

Second, by examining the effects of the model’s circuitry, we have theorized how these interactions feed back into the conditions of human and machinic subjectivity. The metaphors of the Genius Peer, the Matrix, the Loop, and the Stack help shed light on which areas of cognition and empowerment are affected. The Genius Peer represents the risk of false attribution of smartness and reduced critical thought. The Matrix captures the spatial deterritorialization of political agency, where discourse is abstracted from material contexts. The Loop symbolizes the temporal enclosure of imagination, where GenAI systems reinforce existing patterns and foreclose alternative futures. The Stack represents the material entrenchment of technologies into power infrastructures. The effects of the co-en/de model on these conditions are not deterministic, linear outcomes, but contested terrains. Contestation, we argue, shapes the relational subjectivities between humans and GenAI through contrasting processes of Androidization and Cyborgization: The former reduces humans’ autonomy, while the latter augments the possibilities for alternative discourse.

 

The co-en/de model contributes to critical digital and communication studies in four ways. First, it offers a relational ontology of communication that moves beyond techno-determinism and institutionalism. By foregrounding the co-productive nature of human–GenAI interactions, it emphasizes how meaning is negotiated within circuits shaped by power, ideology, and infrastructure. Second, it reframes subjectivity through the processes of Androidization and Cyborgization, allowing us to theorize not only the risks of machinic domination but also the potential for human augmentation and resistance. Third, it integrates spatial and temporal dimensions into the analysis of GenAI communication, situating discourse within broader dynamics of deterritorialization and temporal looping. Finally, it highlights the feedback mechanisms through which micro-level interactions accumulate into macro-level effects, reshaping both the human produser’s agency and the GenAI system’s base of knowledge.

 

Building on these contributions, our theory can inspire a research agenda in three directions. First, future research should empirically investigate the prevalence and effects of the different co-en/de interaction types. Under what conditions do hegemonic interactions dominate? When and how do counter-hegemonic generativity or machinic resistance emerge? Comparative studies across platforms, languages, and political contexts could shed light on how sociotechnical configurations influence these dynamics. Second, future work should explore how human–GenAI interactions intersect with human–human communication. For example, how do GenAI-mediated discourses circulate within social movements, media ecosystems, or institutional settings? Third, we encourage research that explores the infrastructural conditions of resistance. What material and symbolic resources are required to sustain counter-hegemonic interactions? How can grassroots actors intervene in the design, governance, and deployment of GenAI systems to foster emancipatory outcomes?

 

The stakes in human–GenAI interactions are high. As GenAI systems become increasingly integrated into political, social, and cultural life, the circuits of meaning they co-produce shape the contours of democratic agency, public discourse, and collective imagination. The co-en/de model provides a framework for navigating this terrain, not by prescribing solutions, but by illuminating the conditions, dynamics, and possibilities of contestation. Resistance is not guaranteed, nor is enclosure inevitable. The Androidization of the human subject is a tendency, not a destiny. This trend is challenged through critical engagement, collective action, and the reappropriation of technology for public and democratic ends.

 

In this light, the co-en/de model is not only an analytical tool but also a political invitation: to recognize how we are at stake as subjects when interacting with GenAI systems, to interrogate the circuits of meaning we co-produce, and to imagine and enact alternative digitally assisted collective intelligence.

 

 

References

 

Adams, R. (2021). Can artificial intelligence be decolonized? Interdisciplinary Science Reviews, 46(1–2), 176–197. https://doi.org/10.1080/03080188.2020.1840225

 

Alvarado, R. (2023). AI as an epistemic technology. Science and Engineering Ethics, 29(5), Article 32, 1–30. https://doi.org/10.1007/s11948-023-00451-3

 

Arntz, M., Gregory, T., & Zierahn, U. (2017). Revisiting the risk of automation. Economics Letters, 159, 157–161. https://doi.org/10.1016/j.econlet.2017.07.001

 

Bauwens, M., Kostakis, V., & Pazaitis, A. (2019). Peer to peer: The commons manifesto. London, UK: University of Westminster Press. https://doi.org/10.16997/book33

 

Birkinbine, B. J. (2020). Incorporating the digital commons: Corporate involvement in free and open source software. London, UK: University of Westminster Press. https://doi.org/10.16997/book39

 

Bloom, P. (2023). The danger of smart ideologies: Counter-hegemonic intelligence and antagonistic machines. In S. Lindgren (Ed.), Handbook of critical studies of artificial intelligence (pp. 33–42). Cheltenham, UK: Edward Elgar. https://doi.org/10.4337/9781803928562.00008

 

Bødker, H. (2016). Stuart Hall’s encoding/decoding model and the circulation of journalism in the digital landscape. Critical Studies in Media Communication, 33(5), 409–423. https://doi.org/10.1080/15295036.2016.1227862

 

Bonini, T., & Treré, E. (2024). Algorithms of resistance: The everyday fight against platform power. Cambridge, MA: MIT Press.

 

Bratton, B. H. (2016). The stack: On software and sovereignty. Cambridge, MA: MIT Press.

 

Bruns, A. (2013). From prosumption to produsage. In R. Towse & C. Handke (Eds.), Handbook on the digital creative economy (pp. 67–78). Cheltenham, UK: Edward Elgar. https://doi.org/10.4337/9781781004876.00016

 

Bullock, J. B., Chen, Y.-C., Himmelreich, J., Hudson, V. M., Korinek, A., Young, M. M., & Zhang, B. (Eds.). (2024). The Oxford handbook of AI governance. Oxford, UK: Oxford University Press.

 

Burnham, P. (2017). Neo-liberalism, crisis and the contradictions of depoliticisation. Partecipazione e conflitto. The Open Journal of Sociopolitical Studies, 10(2), 357–380. https://doi.org/10.1285/i20356609v10i2p357

 

Chehoudi, R. (2025). Artificial intelligence and democracy: Pathway to progress or decline? Journal of Information Technology & Politics, 1–16. https://doi.org/10.1080/19331681.2025.2473994

 

Coeckelbergh, M. (2024a). Artificial intelligence, the common good, and the democratic deficit in AI governance. AI and Ethics, 5, 1491–1497. https://doi.org/10.1007/s43681-024-00492-9

 

Coeckelbergh, M. (2024b). Why AI undermines democracy and what to do about it. London, UK: Polity Press.

 

Coeckelbergh, M. (2023). Democracy, epistemic agency, and AI: Political epistemology in times of artificial intelligence. AI and Ethics, 3(4), 1341–1350. https://doi.org/10.1007/s43681-022-00239-4

 

Crawford, K. (2021). The atlas of AI: Power, politics, and the planetary costs of artificial intelligence. New Haven, CT: Yale University Press.

 

D’Ignazio, C., & Klein, L. F. (2020). Data feminism. Cambridge, MA: MIT Press.

 

Danaher, J. (2016). The threat of algocracy: Reality, resistance and accommodation. Philosophy & Technology, 29(3), 245–268. https://doi.org/10.1007/s13347-015-0211-1

 

Dattathrani, S., & De’, R. (2023). The concept of agency in the era of artificial intelligence: Dimensions and degrees. Information Systems Frontiers, 25(1), 29–54. https://doi.org/10.1007/s10796-022-10336-8

 

Dijck, J. van (2013). The culture of connectivity: A critical history of social media. Oxford, UK: Oxford University Press.

 

Donoghue, M. (2018). Beyond hegemony: Elaborating on the use of Gramscian concepts in critical discourse analysis for political studies. Political Studies, 66(2), 392–408. https://doi.org/10.1177/0032321717722362

 

Duberry, J. (2022). Artificial intelligence and democracy: Risks and promises of AI-mediated citizen-government relations. Cheltenham, UK: Edward Elgar. https://doi.org/10.4337/9781788977319

 

Dunleavy, P., & Margetts, H. (2023). Data science, artificial intelligence and the third wave of digital era governance. Public Policy and Administration, 40(2), 185–214. https://doi.org/10.1177/09520767231198737

 

Fairclough, N. (2003). Analysing discourse: Textual analysis for social research. London, UK: Routledge.

Filgueiras, F. (2022). The politics of AI: Democracy and authoritarianism in developing countries. Journal of Information Technology & Politics, 19(4), 449–464. https://doi.org/10.1080/19331681.2021.2016543

 

Fuchs, C. (2019). Henri Lefebvre’s theory of the production of space and the critical theory of communication. Communication Theory, 29(2), 129–150. https://doi.org/10.1093/ct/qty025

 

Fuchs, C. (2020). Communication and capitalism: A critical theory. London, UK: University of Westminster Press. https://doi.org/10.16997/book45

 

Fuchs, C. (2023). A Marxist–humanist perspective on Stuart Hall’s communication theory. Theory and Society, 52(6), 995–1029. https://doi.org/10.1007/s11186-023-09524-5

 

Garland, A. (Director). (2014). Ex machina [Film]. University City, CA: Universal Pictures.

 

Gerbaudo, P. (2021). The great recoil: Politics after populism and pandemic. London, UK: Verso Books.

 

Gilbert, J., & Williams, A. (2022). Hegemony now: How big tech and Wall Street won the world (and how we win it back). London, UK: Verso Books.

 

Gillespie, T. (2024). Generative AI and the politics of visibility. Big Data & Society, 11(2), 1–14. https://doi.org/10.1177/20539517241252131

 

González Barman, K., Lohse, S., & de Regt, H. W. (2025). Reinforcement learning from human feedback in LLMs: Whose culture, whose values, whose perspectives? Philosophy & Technology, 38(2), Article 35. https://doi.org/10.1007/s13347-025-00861-0

 

Gramsci, A. (2014). Quaderni dal carcere [Prison notebooks] (V. Gerratana, Ed.). Torino, Italy: Einaudi.

 

Guglielmo, M. (2025). The left and digital politics: Political parties from platform neoliberalism to platform socialism. London, UK: University of Westminster Press. https://doi.org/10.16997/book75

 

Hall, S. (1973). Encoding and decoding in the televisual discourse. Birmingham, UK: University of Birmingham.

 

Hall, S. (1992). Cultural studies and its rheoretical legacies. In L. Grossberg & C. Nelson (Eds.), Cultural studies (pp. 277–294). London, UK: Routledge.

 

Hall, S. (1994). Reflections upon the encoding/decoding model. In J. Cruz & J. Lewis (Eds.), Viewing, reading, listening: Audiences and critical reception (pp. 253–274). Boulder, CO: Westview Press.

 

Hall, S. (2007). Encoding and decoding in the television discourse. In A. Gray, J. Campbell, M. Erickson, S. Hanson, & H. Wood (Eds.), CCCS selected working papers (Vol. 2, pp. 402–414). London, UK: Routledge. https://doi.org/10.4324/9780203357071

 

Hall, S., Nixon, S., & Evans, J. (Eds.). (1997). Representation: Cultural significations and signifying practices. London, UK: SAGE Publications.

 

Haraway, D. (1991). Simians, cyborgs, and women: The reinvention of nature. London, UK: Routledge.

 

Heinrichs, P. S. (2025). Climate change and the imposition of strategic ontologies: Tracing international relations and media studies between militarization, invincibility and vulnerability. Media, War & Conflict, 18(4), 515–531. https://doi.org/10.1177/17506352251350324

 

Herman, E. S., & Chomsky, N. (2011). Manufacturing consent: The political economy of the mass media. New York, NY: Pantheon Books.

 

Jarrett, K. (2015). Feminism, labour and digital media: The digital housewife. London, UK: Routledge.

 

Jarrett, K. (2022). Digital labor. London, UK: Polity Press.

 

Jungherr, A. (2023). Artificial intelligence and democracy: A conceptual framework. Social Media + Society, 9(3), 1–14. https://doi.org/10.1177/20563051231186353

 

Kreps, S., & Kriner, D. L. (2024). The potential impact of emerging technologies on democratic representation: Evidence from a field experiment. New Media & Society, 26(12), 6918–6937. https://doi.org/10.1177/14614448231160526

 

Ku, M. (2025). The digital reproduction of the state: Public diplomacy, digital entitativity, and strategic ontology. Media, War & Conflict, 18(4), 570–595. https://doi.org/10.1177/17506352251350327

 

Kuntsman, A., & Miyake, E. (2022). Paradoxes of digital disengagement: In search of the opt-out button. London, UK: University of Westminster Press.

 

Kwapińska, K. (2022). Technological evolution and the political agency of artificial intelligence from the perspective of general organology and universal organicism. Információs Társadalom, 22(2), 57–71. https://doi.org/10.22503/inftars.XXII.2022.2.4

 

Lerner, A. B., & O’Loughlin, B. (2023). Strategic ontologies: Narrative and meso-level theorizing in international politics. International Studies Quarterly, 67(3), Article sqad058, 1–13. https://doi.org/10.1093/isq/sqad058

 

Lima, E. (2023). The android as a new political subject: The Italian cyberpunk comic ranxerox. In S. Cave & K. Dihal (Eds.), Imagining AI: How the world sees intelligent machines (pp. 55–72). Oxford, UK: Oxford University Press. https://doi.org/10.1093/oso/9780192865366.003.0004

 

List, C. (2021). Group agency and artificial intelligence. Philosophy & Technology, 34(4), 1213–1242. https://doi.org/10.1007/s13347-021-00454-7

 

Manheim, K., & Kaplan, L. (2019). Artificial intelligence: Risks to privacy and democracy. Yale Journal of Law and Technology, 21, 106–188.

 

Mansouri, M. I., & Bailey, D. J. (2025). How to be “anti-AI” in the 21st century: Overcoming the inevitability narrative. Global Political Economy, 4(2), 185–194. https://doi.org/10.1332/26352257y2025d000000030

 

Mazzucato, M., Schaake, M., Krier, S., & Entsminger, J. (2022). Governing artificial intelligence in the public interest (Vol. 2). London, UK: UCL Institute for Innovation and Public Purpose. https://www.ucl.ac.uk/bartlett/public-purpose/wp2022-12

 

McDowell, J. C. (2014). The politics of big fantasy: The ideologies of Star Wars, The Matrix and The Avengers. Jefferson, NC: McFarland.

 

Michelle, C. (2007). Modes of reception: A consolidated analytical framework. The Communication Review, 10(3), 181–222. https://doi.org/10.1080/10714420701528057

 

Miragoli, M. (2025). Conformism, ignorance & injustice: AI as a tool of epistemic oppression. Episteme, 22(2), 522–540. https://doi.org/10.1017/epi.2024.11

 

Moore, P., & Woodcock, J. (Eds.). (2021). Augmented exploitation: Artificial intelligence, automation and work. London, UK: Pluto Press.

 

Mühlhoff, R. (2025). The ethics of AI: Power, critique, responsibility. Bristol, UK: Bristol University Press.

 

Muldoon, J., Graham, M., & Cant, C. (2024). Feeding the machine: The hidden human labour powering AI. Edinburgh, UK: Canongate Books.

 

O’Loughlin, B., Guglielmo, M., McInerney, L., & Pearson, E. (2025). AI and political agency: Learning to live together. Political Insight, 16(1), 29–31. https://doi.org/10.1177/20419058251332343

 

Ott, B. L., & Mack, R. L. (2013). Critical media studies: An introduction. Hoboken, NJ: John Wiley & Sons.

 

Paul, R., Carmel, E., & Cobbe, J. (Eds.). (2024). Handbook on public policy and artificial intelligence. Cheltenham, UK: Edward Elgar. https://doi.org/10.4337/9781803922171.00006

 

Petley, J. (2022). “Well grubbed, old mole!”: The press, the Institute of Economic Affairs and the propagation of neo-liberalism in the UK. Journalism, 23(9), 1864–1880. https://doi.org/10.1177/14648849211015853

 

Radu, R. (2021). Steering the governance of artificial intelligence: National strategies in perspective. Policy and Society, 40(2), 178–193. https://doi.org/10.1080/14494035.2021.1929728

 

Ronge, R., Maier, M., & Rathgeber, B. (2025). Towards a definition of generative artificial intelligence. Philosophy & Technology, 38(1), 1–25. https://doi.org/10.1007/s13347-025-00863-y

 

Ross, S. (2011, May). The encoding/decoding model revisited. Paper presented at Annual Meeting of the International Communication Association, Boston, MA. https://www.researchgate.net/profile/Sven-Ross/publication/346010378_THE_ENCODINGDECODING_MODEL_REVISITED/links/5fb58b41a6fdcc6cc649dd22/THE-ENCODING-DECODING-MODEL-REVISITED.pdf

 

Schmidhuber, J., Schlögl, S., & Ploder, C. (2021). Cognitive load and productivity implications in human–chatbot interaction. In 2021 IEEE 2nd International Conference on Human–Machine Systems (ICHMS) (pp. 1–6). Magdeburg, Germany: ICHMS. https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9582445&casa_token=Sejx7flHfdcAAAAA:s7c0yTrDh8y6B-OXJozC1i10IaAKTyN45ClYWCRhTQqJoRo-GG3wGJ3b84qGyDjrhdh675BB&tag=1

 

Scott, R. (Director). (1982). Blade runner [Film]. Burbank, CA: Warner Bros.

 

Supriyanto, E. E., & Saputra, J. (2022). Big data and artificial intelligence in policy making: A mini-review approach. International Journal of Advances in Social Sciences and Humanities, 1(2), 58–65. https://doi.org/10.56225/ijassh.v1i2.40

 

Treré, E. (2018). Hybrid media activism: Ecologies, imaginaries, algorithms. London, UK: Routledge.

 

Veale, M., & Zuiderveen Borgesius, F. (2021). Demystifying the draft EU Artificial Intelligence Act—Analysing the good, the bad, and the unclear elements of the proposed approach. Computer Law Review International, 22(4), 97–112. https://doi.org/10.3390/philosophies10050095

 

Vera Hoyos, C., & Cárdenas Marín, W. (2025). The use of artificial intelligence in political decision-making. Philosophies, 10(5), 95, 1–44. https://doi.org/10.3390/philosophies10050095

 

Wachowski, A., & Wachowski, L. (Directors). (1999). The matrix [Film]. Burbank, CA: Warner Home Video.

 

Wang, P. (2019). On defining artificial intelligence. Journal of Artificial General Intelligence, 10(2), 1–37. https://doi.org/10.2478/jagi-2019-0002

 

Williams, A. (2019). Political hegemony and social complexity: Mechanisms of power after Gramsci. London, UK: Palgrave Macmillan.

 

Williams, R. (2016). Communications. London, UK: Random House. (Original work published 1962)

 

Wodak, R. (2015). The politics of fear: What right-wing populist discourses mean. London, UK: SAGE Publications.

 

Worth, O. (2015). Rethinking hegemony. London, UK: Palgrave Macmillan.

 

 

Copyright © 2026 (Marco Guglielmo, Ben O’Loughlin, and Henry Snowball).

Licensed under the Creative Commons Attribution Non-commercial No Derivatives (by-nc-nd).

https://doi.org/10.65476/1sxwvg58

 

[1] Acknowledgements: We warmly thank the participants and our co-organizers (Dr. Elizabeth Pearson and Dr. Lisa McInerney) for the workshop “Political Agency Between Humans and Artificial Intelligence” (London, 12/09/2024), which provided key insights for the development of this study. Thanks, too, to the peer reviewers and editorial team.

[2] Statement of Funding: Marco Guglielmo’s research is funded by the European Union, Grant: No. Horizon-MSCA-2022-PF-01-101110457. Views and opinions expressed are those of the author only.