The intersection of dossierveillance with social media propaganda

Not like overt surveillance

Surveillance is ubiquitous in modern society. As Humphreys (2011) explains, “information technology and new media allow for collecting and sharing personal information at unprecedented levels” (p. 575). In contrast to overt surveillance, however, dossierveillance does not require active monitoring to induce behavioral change. Instead, it relies on the internalization of the surveillant gaze. Drawing from Foucauldian notions of the panopticon, dossierveillance exists in the minds of individuals who fear that a bureaucratic record – real or imagined – may be assembled against them, thereby impacting their reputation, employment, freedom, or safety. Yet, in line with Foucault’s (1982, 1988) later work on technologies of the self, individuals are not wholly passive: they adapt, negotiate, and even creatively leverage these conditions to express dissent, cultivate alternative identities, or engage in subtle forms of resistance. It introduces a psychological structure of self-restraint, where individuals regulate speech, conduct, and even beliefs due to the haunting presence of potential documentation (Plamadeala, 2019b; 2025a). Dossierveillance is often invisible and undefined, unlike CCTV cameras or visible online tracking (see work by Alikhanov and Kim, 2023). Rather than functioning solely through coercion, its power also stems from uncertainty, ambiguity, and the anticipation that one’s actions may later be interpreted, archived, or weaponized by institutional actors.

The dynamics of dossierveillance become particularly tangible in digital workplace controversies. As we will see later, in the James Damore case, for instance, an internal Google memo intended for limited circulation became a widely disseminated reputational dossier after being leaked online. The controversy shows how institutional communication within platform societies can rapidly transform into a searchable public archive carrying professional consequences (Kunzru, 2022).

In like fashion, employees may refrain from posting on social media about workplace conditions for fear their opinions could be misinterpreted or saved in a digital dossier. However, others might adopt pseudonymous accounts or private forums to maintain agency over expression. Students may censor their classroom contributions if they believe recordings or emails could be archived and weaponized. The threat need not be actualized – only imagined – for the behavioral modification to occur (Plamadeala, 2025a; 2025b).

Centralized vs. institutional forms

Fundamental to the theory is the distinction between centralized and institutional forms. Centralized dossierveillance is typified by totalitarian régimes like Ceauşescu’s Romania or the GDR, where the Securitate or Stasi systematically developed bureaucratic files on citizens through hierarchical command structures and grassroots informants (Plamadeala 2025a). In contrast, institutional dossierveillance, more prevalent in contemporary neo-liberal societies like those in the Western world, manifests through decentralized yet pervasive mechanisms within schools, workplaces, health care systems, and digital platforms. Crucially, these mechanisms allow for both constraint and strategic agency, as individuals interpret and respond to potential surveillance in diverse ways.

While centralized forms often involve top-down orchestration by state intelligence services, institutional dossierveillance operates laterally, enforced not by secret police but by the everyday bureaucracies and platforms we interact with. This lateral form of surveillance is sometimes explained through concepts such as lateral surveillance, peer-to-peer monitoring, or situations where users willingly accept institutional oversight, which helps understand these dynamics within contemporary digital practices (Andrejevic, 2004). Plamadeala (2025a) also adds the “deliberant” category to differentiate between centralized dossierveillance experienced in totalitarian societies and that in current neo-liberal ones. This distinction becomes clearer when comparing Cold War surveillance systems with contemporary platform infrastructures. Whereas the Stasi relied on centralized physical archives compiled by informants (Gieseke, 2014), modern institutional dossierveillance emerges through dispersed digital systems that continuously aggregate metadata, engagement histories, screenshots, and behavioral analytics (Plamadeala, 2025a).

The word “deliberant” refers to the intentional harm imposed by governmental authorities such as the Securitate or the Stasi police in Communist Romania or the GDR. The campaigns and strategies from these régimes sought to deliberately and self-evidently harm their targets. By the same token, in contemporary surveillance societies, most institutional dossierveillance is of non-deliberate type. Plamadeala (2025a) also introduces another category for dossierveillance, referred to as “multi-layered,” wherein multiple surveillant “gazes” are involved, including those within an institution (Foucault, 1977). This study primarily focuses on these latter two forms, emphasizing their interaction with digital platforms and peer surveillance, while acknowledging that individuals retain capacities for resistance and self-fashioning.

Institutional dossierveillance emerges when universities monitor student emails, when medical systems maintain behavioral health records, or when social media platforms archive user interactions and facial recognition data. The cumulative effect is a structure of soft control: not through fear of prison, but fear of reputation damage, job loss, or cancellation (Beetham et al., 2022). Nevertheless, empirical studies suggest that many individuals respond selectively, finding ways to protect privacy or challenge perceived threats, highlighting the non-deterministic nature of modern dossierveillance. As such, in a study conducted by Ioannou and Tussyadiah (2021), it was revealed that, although individuals may accept surveillance in some contexts (e.g., for public health reasons), many still engage in active privacy protection behaviors (e.g., reducing the amount of personal data shared online or changing privacy settings) to mitigate intrusion, demonstrating that responses are conditional and strategic rather than uniformly accepting.

Weaponizing emotion and identity in the digital age

One of the most effective mechanisms of social media propaganda is its ability to exploit users’ emotions and identities. Emotional content – particularly fear, outrage, pride, and resentment – travels faster and farther than factual reporting. Propagandists understand this and tailor messages to trigger strong affective responses (Horner et al., 2021; Mattingly and Yao, 2022).

The operations of the Internet Research Agency (IRA) during the 2016 U.S. election illustrate how emotional manipulation and dossierveillance converge in practice. IRA-linked accounts impersonated activists, infiltrated online communities, and amplified emotionally polarizing narratives surrounding race, policing, and nationalism. These campaigns did not simply attempt to persuade audiences ideologically; they generated emotionally charged interactions that could later be algorithmically categorized, monitored, and strategically targeted (Cosentino, 2020).

A study by MIT in 2018 found that false news stories are 70% more likely to be retweeted than true ones, largely because they evoke surprise or fear. Platforms’ engagement-driven algorithms, which prioritize emotionally resonant content, inadvertently reward propaganda over neutrality (Vosoughi et al., 2018). This emotional manipulation often intersects with identity politics. Propagandists craft messages that affirm the beliefs of specific communities or social groups – whether religious, political, ethnic, or cultural – and vilify perceived outsiders. Yet, empirical studies indicate that users sometimes respond with skepticism, humor, or counter-narratives, illustrating the heterogeneity of outcomes.

What makes emotional propaganda particularly insidious is its ability to mobilize collective action – online and offline. Hashtag campaigns, viral videos, and coordinated disinformation can provoke protests, unrest, or even violence. The 2021 Capitol insurrection in the U.S. was partly fueled by months of digital propaganda on platforms like Twitter, YouTube, and Parler, where narratives of a “stolen election” were seeded and nurtured by both grassroots users and elite influencers (Matusitz, 2022). In like fashion, in countries like Myanmar and India, inflammatory social media rhetoric has been linked to outbreaks of ethnic violence and vigilantism. Furthermore, emerging studies suggest that algorithmically amplified content often accelerates these dynamics by prioritizing emotionally charged narratives over factual accuracy, creating feedback loops that magnify public outrage (Cinelli et al., 2022; Murero, 2023).

In this context, it is important to explicitly address the concept of “affordances,” which refers to the action possibilities that social media platforms inherently provide to users and propagandists. Algorithmic structures, content moderation practices, and engagement mechanisms constitute affordances that shape how users produce, share, and respond to emotional or identity-based content. Recognizing affordances allows us to see how platform design, rather than solely user intention, structures the amplification of propaganda and the mobilization of collective behavior, making certain manipulative strategies technically feasible and socially consequential (Ronzhyn et al., 2022).

In this emotional economy, truth becomes secondary to feelings, and propaganda thrives not by convincing the rational mind but by capturing the emotional body. The architecture of social media – real-time updates, echo chambers, algorithmic amplification – creates an environment where people are not only more susceptible to propaganda but are primed to spread it. In such a climate, content that affirms users’ fears and identities holds greater currency than fact-checked journalism, opening the door for propagandists to weaponize the very emotions that bind communities together (Forest, 2021; Woolley, 2023). This dynamic has been particularly evident in fast-moving crises, where immediate emotional reactions can outpace fact-checking, granting early sharers disproportionate influence over public perception.

Coordinated inauthentic behavior (CIB)

While propaganda has always involved a degree of manipulation, the digital age has introduced new levels of sophistication, scale, and automation. The distinction between organic influence and orchestrated campaigns has blurred, creating environments in which users cannot reliably distinguish authentic peer discourse from engineered persuasion. Propaganda infrastructures simultaneously function as systems of dossier accumulation. Bot networks do not merely amplify content; they generate data about which narratives produce outrage, fear, loyalty, or engagement. Through continuous interaction, these systems refine user profiles and emotional categorizations, transforming political participation into behavioral metadata. In this sense, coordinated inauthentic behavior is not only propagandistic but archival: every interaction becomes part of a continuously evolving digital dossier (Woolley and Howard, 2018).

A key feature of modern social media propaganda is “coordinated inauthentic behavior” (CIB), a term used by Facebook (now Meta) to describe efforts where groups of accounts work together to mislead users about their identity, intent, or source of information (Murero, 2023). These campaigns are often designed to appear organic and grassroots but are, in fact, centrally orchestrated – sometimes by foreign governments, political consultants, or ideological movements (Cinelli et al., 2022; Murero, 2023). In contemporary operations, AI-generated accounts, deepfake personas, and synthetic media are deployed to simulate human behavior, thereby expanding reach while reducing detection risk, illustrating the convergence of machine learning and psychological manipulation in CIB.

One of the most striking examples of CIB was the operation run by Cambridge Analytica, a data analytics firm that harvested personal data from millions of Facebook users without their consent to create psychographic profiles. These profiles were then used to micro-target users with tailored political advertisements during the Brexit referendum and the 2016 U.S. presidential election. Although the firm was eventually exposed and dissolved, the episode demonstrated how behavioral data could be weaponized to influence individual voters with surgical precision (Dmytro and Viacheslav, 2025; Kreiss and Barrett, 2020). Beyond the ethical lapses, this case exemplifies a broader structural problem: political influence now operates through invisible architectures of affective targeting, where persuasion is measured not by public debate but by the precision of emotional nudges. The exposure of such operations underscores the blurred line between legitimate data analytics and manipulative propaganda, highlighting ethical and regulatory gaps.

In other cases, state actors have employed armies of bots – automated accounts programmed to like, comment, retweet, and share content – to simulate popularity or legitimacy. For instance, investigations into the Kremlin-backed disinformation ecosystem revealed that thousands of Twitter bots amplified divisive hashtags, such as #ReleaseTheMemo or #Calexit, designed to sow discord in the U.S. political system (PBS NewsHour, 2018). Similarly, during the COVID-19 pandemic, bot-driven networks were used to spread anti-vaccine misinformation and conspiracy theories, thereby undermining public health efforts (Unlu et al., 2024). These networks demonstrate that algorithmic virality can outpace fact-checking, creating affective cascades where emotion-driven content gains authority irrespective of truth, reshaping both public opinion and institutional trust. These campaigns often exploit algorithmic biases that favor virality, making automated amplification more effective than organic sharing alone.

These coordinated operations often include influencers, fake news sites, and trolls working in tandem to manipulate platform trends and public perception. They exploit platform algorithms, timing, and even search engine optimization to ensure their content is highly visible. In some cases, propaganda farms flood platforms with content not to persuade but to overwhelm, diluting credible information in a sea of contradictory messages – a tactic known as “cognitive overload” (Marjanović and Smiljanić, 2025, p. 95) or “flooding the zone” (Bince, 2024, p. 169). Cognitive overload, in particular, compounds the pressure on social media users to post continuous content, usually to the detriment of truth or accuracy (Joseph and Rickett, 2014). This environment transforms engagement from a voluntary act into an affective labor, where users unconsciously participate in propagandistic cycles, reinforcing the power of automated systems and amplifying socio-psychological vulnerabilities. It should also be emphasized that cognitive overload disproportionately affects less experienced users, who are less able to critically evaluate information, thus amplifying the social impact of these campaigns.

Because these campaigns often mimic organic user behavior, they are difficult to detect, particularly when the lines between state actors, private contractors, and ideological volunteers blur. The result is an environment where inauthentic actors shape authentic belief, and users are left unable to discern who is speaking, why they are speaking, or whether the conversation was real to begin with. Social media, thus, becomes not merely a battlefield of ideas but a theater of manufactured consensus (Woolley, 2023). This manufactured consensus functions affectively, creating the perception of majority opinion, which in turn pressures dissenting voices to self-censor, internalizing control mechanisms without overt enforcement. In due course, it can erode trust in media institutions and public deliberation, further destabilizing civic life.

Social media propaganda and the reanimation of dossierveillance

Contemporary social media propaganda, particularly as wielded by state actors, ideological extremists, or political operatives, operates not only through data collection but through the cultivation of ambient anxieties, producing an ecosystem in which users internalize the threat of exposure, cancellation, or reputational harm. The reanimation of dossierveillance becomes concrete in platform controversies where digital traces are transformed into instruments of social discipline. During political crises or ideological conflicts, users observe how screenshots, archived posts, leaked emails, and historical comments can rapidly circulate beyond their original contexts. Witnessing these consequences encourages anticipatory self-censorship even among uninvolved observers. Thus, propaganda succeeds not merely through persuasion but through the public demonstration that digital participation leaves retrievable traces capable of future mobilization. In his article in Communication Theory, Soffer (2021) gives details on how algorithmic personalization follows the two-step flow of communication, substituting traditional opinion leaders with digital curators who guide the propagation of content through influencer networks while simultaneously performing a continuous, invisible gatekeeping function.

At the heart of this intersection is the production of “data doubles” – synthetic digital personas constructed from a user’s behaviors, interactions, language patterns, network connections, and engagement histories (Bridges, 2021). These data doubles allow propagandists to micro-target users with ideologically charged content, disinformation, or consumer messaging, transforming historical surveillance dossiers into continuously evolving algorithmic scripts that operate invisibly, perpetually updating, and adapting to perceived vulnerabilities (Huq, 2021). While historical dossiers were compiled by centralized authorities and stored physically, modern data doubles are aggregated, inferred, and circulated automatically, often without awareness or consent, producing what could be described as a living archive of affective and behavioral surveillance.

More insidiously, social media propaganda leverages affective engineering to induce self-censorship, a byproduct of dossierveillance. Plamadeala (2025a) observes that dossierveillance encompasses both tangible surveillance artifacts – files, digital footprints, and algorithmically derived dossiers – and an affective dimension, the subjective experience of being observed, monitored, or evaluated. The affective dimension is weaponized through deliberate emotional design: campaigns evoke fear, shame, outrage, or guilt, prompting users to adjust their behavior preemptively. Erickson and Yan (2024) demonstrate that these manipulations are calibrated to amplify emotional responses while reinforcing ideological positions.

Into the bargain, the presence of peers, influencers, and anonymous audiences amplifies this effect. Individuals witnessing online harassment, doxing, or reputational attacks against others may feel compelled to moderate their political speech or disengage from controversial topics altogether. This reflects Plamadeala’s (2025a) notion of multi-layered, non-deliberate dossierveillance, wherein the mere possibility of digital documentation induces internalized restraint. Many propagandistic campaigns rely not only on informational manipulation but also through the systematic cultivation of suspicion, mistrust, and affective insecurity, echoing the paranoia historically produced by authoritarian intelligence apparatuses (Nasiri and Hashemzadeh, 2025).

Consequently, social media propaganda does not simply disseminate ideology; it reanimates the bureaucratic logic of dossierveillance in decentralized digital networks. Platforms become arenas in which users are simultaneously observers and observed, curators and curated, producing horizontal surveillance loops where peer policing, algorithmic reinforcement, and self-surveillance converge. This environment mirrors historical strategies, such as those of the Securitate, but in a digitally networked, emotionally charged, and continuously replicating form (Plamadeala, 2025a).

The emotional mechanics of dossierveillance in platform societies

A key contribution of the dossierveillance framework lies in its emphasis on the emotional and psychological experience of being surveilled, particularly within liberal-democratic and capitalist contexts where surveillance may not appear overtly coercive. The Kevin Hart controversy attests to the fact that emotional self-regulation emerges through archival visibility. Hart’s decade-old tweets were resurfaced and recirculated during public debate over his appointment as Oscars host. Regardless of one’s normative evaluation of the controversy, the case reveals how digital archives create anticipatory anxiety for users who recognize that historical expressions may later be detached from their original temporal and cultural contexts (Dawson et al., 2023).

Understanding these dynamics benefits from an awareness of media logic and platformization, which highlight how platforms’ connective design, algorithmic governance, and engagement incentives condition both fear and empowerment (Altheide, 2015; Van Dijck et al., 2018). In current platform societies, the boundary between state surveillance, corporate tracking, and peer scrutiny has collapsed. Users no longer fear only the state; they also fear employers, social networks, and public backlash. This fear operates as an invisible vector of compliance, reinforcing platform governance through affective conditioning rather than formal enforcement mechanisms (Kruse et al., 2018). This is especially evident in cancel culture, workplace surveillance, and algorithmic moderation.

With this said, lateral forms of surveillance mediated by peers can also produce feelings of agency, for instance when users actively monitor, curate, or manage their own digital presence to achieve social or political goals, demonstrating that dossierveillance is not uniformly coercive (Koskela, 2004; McGrath, 2004). For instance, a whistleblower contemplating whether to expose unethical practices within their company may hesitate, not because of direct threats, but due to the possibility of being flagged, blacklisted, or misrepresented online. This hesitation embodies the affective dimension of dossierveillance: the knowledge or suspicion that a file – whether digital or social – may exist or be created against them. It also shows how dataveillance (passive data collection) can morph into dossierveillance (emotional surveillance) when the user begins to self-monitor due to anticipated creation of a digital dossier that could be weaponized against them, demonstrating the affective dimension of dossierveillance. Dataveillance, defined by Clarke (1994) as passive, systematic data monitoring, transforms into dossierveillance once users internalize the risk of this data being mobilized for coercion or judgment.

In line with these contentions, public figures and ordinary users alike frequently regulate their speech in anticipation of being screen-captured, decontextualized, and archived. In these instances, the “dossier” is not a physical file, but a distributed cultural artifact – tweets, recordings, posts – that could be mobilized against them in contexts they cannot predict. Such anticipatory self-regulation reproduces coercive dynamics reminiscent of totalitarian surveillance, fostering fear, paralysis, and conformity without centralized enforcement.

Targeted propaganda campaigns exploit this dynamic by signaling that one’s affiliations or behaviors are being continuously monitored and may be judged publicly, a form of psychographic manipulation that weaponizes identity-based affective responses (Plamadeala, 2019b). These campaigns leverage insecurities and social fault lines, cultivating what Hannah Arendt (1968) described as psychological readiness to obey or conform out of fear, rather than coercion alone.

In effect, dossierveillance has become part of the emotional infrastructure of social media life. Even when no actual data collection is occurring, the belief that one is being watched, that a trace of dissent could jeopardize one’s status, renders the phenomenon operative. Thus, the power of dossierveillance derives less from technical accuracy than from its capacity to mobilize fear, mistrust, shame, and anxiety as instruments of social control.

A tool for diagnosing surveillance societies

Given the rise of surveillance capitalism and the intensification of ideological propaganda online, dossierveillance offers an urgently needed framework for diagnosing how fear functions as a governing mechanism alongside law, policy, and technology. With surveillance capitalism, behavioral data is continuously extracted and monetized for predictive control (Zuboff, 2019). Similarly, dossierveillance is not limited to documenting behavior; it operates as a form of governance by projecting future outcomes, shaping emotions, and cultivating anxiety about potential exposure. These dynamics are intensified by context collapse, where diverse audiences converge in the same digital space, increasing the risk of misinterpretation and reputational collapse (Marwick & boyd, 2011).

The Cambridge Analytica scandal demonstrates how behavioral dossiers can be operationalized for political micro-targeting; the Damore controversy illustrates how institutional communication becomes reputational surveillance; IRA disinformation campaigns reveal how emotional polarization generates behavioral archives; and cancel-culture controversies such as the Kevin Hart case show how digital traces are retrospectively weaponized within affective publics. Collectively, these examples reveal that dossierveillance is not an isolated phenomenon but a recurring structural condition of platform societies.

However, integrating insights from platform studies reminds us that the emotional impact of surveillance is mediated by both algorithmic design and peer networks, which can sometimes enable user empowerment or lateral visibility. Nevertheless, this framework does not presume that fear is the only or even dominant affective response to surveillance, nor that subjects are uniformly disciplined by it. Rather, dossierveillance is proposed as one analytic lens among others for examining how surveillance becomes emotionally meaningful under specific political, cultural, and technological conditions. Whereas dataveillance provides an analytical model for understanding the systems that collect and store personal data (Clarke, 1994), dossierveillance focuses on the human response to these systems. It interrogates how individuals are made to feel vulnerable, observed, and categorized – often in ways that stifle dissent and foster complicity (see Plamadeala 2025a), but also, in some contexts, how they negotiate, reinterpret, resist, or strategically accommodate these conditions through practices of self-fashioning, irony, silence, or selective disclosure.

Dossierveillance also enables scholars and critics to move beyond simplistic binaries of authoritarian vs. democratic or East vs. West. By identifying common affective and institutional mechanisms across different régimes, it highlights how even ostensibly free societies can reproduce authoritarian traits through bureaucratic record-keeping, algorithmic classification, and digital ostracism. Yet, this comparative move does not imply equivalence between political systems, nor does it deny the presence of countervailing forces such as legal protections, journalistic accountability, civic resistance, and everyday forms of agency. In this sense, dossierveillance bridges historical surveillance studies with contemporary digital media research, creating a continuum that reveals the persistence of certain coercive logics. In Surveillance & Society, Volinz (2025) critiques the authoritarian-democratic binary by revealing how routine bureaucratic surveillance, algorithmic classification, and institutional ostracism span diverse political systems, while also showing how these practices are unevenly contested and politically negotiated rather than mechanically imposed.

Moreover, the theory’s distinction between deliberate and non-deliberate forms of dossierveillance is especially valuable for understanding the gray areas of social media propaganda. Not all actors engaging in data collection or behavioral targeting are driven by conscious intent to harm. Yet, the effects – fear, silence, retreat – mirror those produced by malicious surveillance. Part of the reason is that surveillance generates, on occasion, apathy, resignation, indifference, strategic compliance, or even playful appropriation, complicating any singular account of fear as the dominant outcome. To this point, Ullrich and Knopp (2018) examined how activists adapt to and respond ambiguously to police video surveillance through a variety of tactics (such as disguise, sousveillance, and cooperation) showing that subjects are not simply deterred by monitoring but adjust, resist, or appropriate the surveillant gaze in practice. This obfuscates moral and legal evaluations and suggests that surveillance harm must also be assessed from the standpoint of the perceived impact on the subject, not just the intent of the surveillant (Plamadeala 2025a).

From an ethical and political standpoint, Plamadeala (2025a) applies dossierveillance to raise fundamental questions about agency, democracy, and digital citizenship. How can individuals meaningfully participate in public discourse when they fear being archived, judged, or punished? Equally important, how do individuals actively craft digital selves, experiment with irony, cultivate ambiguity, or withdraw strategically in ways that complicate purely disciplinary readings of surveillance power? What does freedom of speech mean in a context where every utterance may be scraped, categorized, and misinterpreted? And what responsibilities do institutions – be they states, platforms, or employers – bear in curbing the atmospheres of fear and mistrust that dossierveillance cultivates?

In response, scholars and activists may use the framework of dossierveillance to advocate for more transparent data governance, stronger protections for whistleblowers, and media literacy programs that demystify the ways in which digital profiles are used to shape behavior. They might also push for platform architectures that discourage peer-policing and harassment while promoting spaces for genuine, dissenting dialogue without fear of bureaucratic retaliation. Taken together, these interventions underscore that dossierveillance does not describe an inescapable regime of domination, but rather a contingent configuration of power that remains open to contestation, reinterpretation, and institutional reform. In the end, dossierveillance illuminates the emotional and bureaucratic scaffolding of modern propaganda. It reminds us that surveillance is not just a technological condition, but a psychological and political one. And in so doing, it offers a lens for reclaiming our digital selves from the bureaucratic shadows we are taught to fear.

Affective publics in propaganda networks

While the other three factors discussed in this section focus on institutional, ideological, psychological, and platform-level intersections between dossierveillance and propaganda, an equally vital yet underexplored terrain is that of affective publics – emotionally charged online communities mobilized by shared feelings rather than rational consensus (Papacharissi, 2015). Affective publics become especially important in understanding how ordinary users participate in the production of dossiers. Viral outrage cycles, screenshot cultures, reposting practices, and call-out campaigns all transform emotional participation into searchable archives. In these environments, users are not merely surveilled from above by institutions; they also surveil one another horizontally through continuous acts of observation, documentation, and recirculation (Bucher, 2018).

These publics are shaped not only by affective engagement but also by platform logics that govern visibility, attention, and data aggregation, further influencing how users perceive risk, reward, and empowerment online (Altheide, 2015; Koskela, 2004; McGrath, 2004; Van Dijck et al., 2018). Equally important, these publics cannot be reduced to passive recipients of propaganda: they are also sites of connective action, creative expression, irony, and political experimentation, where users actively shape meanings rather than merely absorb them. This emotional infrastructure, as cultivated through social media propaganda, becomes a fertile ground for the emergence of dossierveillance as an internalized mode of control. In affective publics, emotion is not simply a reaction to content but becomes a currency of exchange and a vector of political mobilization. Here, dossierveillance operates not through bureaucratic dossiers in the classical sense, but through affective traces: likes, shares, reactions, deleted posts, and screenshots – micro-records embedded in the social fabric of online interaction.

Social media propaganda functions by generating strong emotional responses – outrage, indignation, patriotism, fear – and by encouraging users to express these feelings publicly. Nonetheless, these emotional expressions frequently function not only as propaganda but also as connective practices that build solidarity, humor, critique, and political belonging, particularly through vernacular genres such as memes, remixes, and ironic commentary. However, these very expressions are then stored, algorithmically categorized, and made retrievable in ways that can later be used to surveil, discipline, or discredit. The individual who participates in affective publics may later self-censor, worrying about how a past tweet expressing anger, criticism, or political dissent may be decontextualized and weaponized (Woolley, 2023). This dynamic does not imply that fear is universal or constant; rather, it highlights one pathway through which emotional participation may become retrospectively reinterpreted as risk. This fear, based on real or imagined records of one’s emotional expressions, is a quintessential instance of dossierveillance: the internalization of the surveillant gaze through the emotionally encoded data one has left behind.

Of equal relevance is the fact that this affective dossierveillance is not always orchestrated by the state or a central authority. Instead, it is often perpetuated horizontally by other users, followers, or algorithmic agents who resurface past behaviors or expressions. The emotional economy of social media thus creates a form of dossierveillance that is peer-driven, ambient, and decentralized, but no less disciplining. Cancel culture, call-out posts, and screenshot exposés are all examples where affective expressions are reassembled into dossiers that can incite moral panic or reputational damage (see Bouvier and Machin, 2021; Grunsven and Marin, 2024). These practices can even function as mechanisms of accountability, community norm-setting, and collective critique, underscoring the ambivalence of affective surveillance as both a tool of discipline and a resource for democratic contestation.

Propaganda agents exploit this dynamic by designing campaigns that intentionally provoke affective responses while simultaneously constructing digital environments where those responses are meticulously recorded. Thus, people are drawn into cycles of reaction, participation, and self-censorship. In this ecosystem, dossierveillance serves as both the architecture of memory and the mechanism of control. The archive of past emotional expression becomes both a badge of belonging and a source of anxiety. The line between participation and exposure blurs, producing a social media subject who is not only a propagandized agent but also a curator of their own dossier, always aware of its potential future consequences. In essence, this subject is not merely disciplined but also reflexive, improvisational, and capable of tactical opacity, irony, and withdrawal, complicating any totalizing account of affective control.

Normalization of self-censorship

In the 21st century, the convergence of dossierveillance and social media propaganda has fundamentally reshaped the public’s relationship with free expression. The James Damore controversy reveals how public examples of reputational punishment normalize anticipatory self-censorship within professional and institutional settings. Employees witnessing the backlash surrounding leaked internal communication may subsequently avoid expressing dissenting views, not because explicit censorship exists, but because digital visibility creates uncertainty regarding future consequences (Kunzru, 2022).

What was once regarded as a space for open dialogue and political dissent is now increasingly structured by fear, suspicion, and self-restraint. Propaganda does not merely deceive or mislead – it conditions people to become cautious in what they post, like, or share. This results in an internalization of the surveillant gaze: individuals constantly manage their digital personas, not only to avoid backlash from other users but also to escape algorithmic scrutiny or institutional retribution. Lyon’s (2017) concept of a surveillance culture helps explain how such internalization becomes normalized, embedding expectations of monitoring into everyday social and professional practices. Dossierveillance contributes to this normalization of self-censorship by converting every interaction into a potential entry in an unseen dossier – one that may be compiled, interpreted, and mobilized in unforeseeable ways.

As proof that a certain percentage of users attempt to escape algorithmic scrutiny or institutional retribution, a study conducted by Weeks et al. (2024) showed that fear of social sanctions (FOSS) – concerns about personal/professional repercussions and social isolation from one’s network – is a measurable predictor of reduced political expression on social media. People embedded in socially diverse networks were more likely to self-censor due to fears of negative reactions from others. This supports the idea that fear of backlash from people (friends, family, coworkers) strongly shapes online self-presentation and expression.

What sets this apart from traditional censorship is its subtlety; the fear is not always of the state, but of being misunderstood, misrepresented, or misjudged in the future. It is a fear of one’s digital trace being recontextualized in ways that carry reputational or professional consequences. On the other hand, empirical work also shows that social media can amplify incivility and polarization, intensifying criticism and confrontation rather than uniformly silencing users. For instance, Brundidge and Garrett’s (2024) longitudinal study empirically demonstrates that exposure to incivility on social media interacts with online news use to intensify affective polarization, showing how uncivil and antagonistic discourse on platforms contributes to deeper political divisions rather than dampening criticism.

An interesting example is the case of James Damore, the former Google engineer who, in 2017, wrote an internal memo criticizing aspects of Google’s diversity policies (Kunzru, 2022). Though intended for internal discussion, the document quickly circulated online and was labeled by critics as sexist and discriminatory. The backlash was swift – Damore was fired, and his memo became the focal point of a broader public debate about free speech, ideological conformity, and the limits of dissent in corporate environments (Bermiss and McDonald, 2018). Regardless of one’s stance on the content of his memo, the case exemplifies the mechanisms of non-deliberate institutional dossierveillance: internal records, intended as private communication, became the basis for public condemnation and employment consequences. The dossier here was both literal (the memo itself) and affective (the media firestorm that followed), combining to chill future expression within and beyond Google. Moreover, comparative research indicates that social media users encounter more visible political disagreement online than in face-to-face contexts, often leading to heightened intensity of critique and outspoken engagement. The Barnidge (2017) article finds that users perceive more political disagreement on social media compared with face-to-face interaction and that such disagreement on social platforms can be more visible and intense, supporting the claim that online contexts often facilitate outspoken and confrontational critique relative to offline discourse.

This climate encourages people to preemptively self-regulate, even in the absence of clear rules. Dissent, sarcasm, or controversial opinions are quietly omitted. For many, this is not merely prudence but survival – a way to navigate professional or social spaces that now overlap with surveillance régimes. Thus, social media propaganda, reinforced by dossierveillance, does not need to silence everyone explicitly; it fosters a culture where silence and neutrality feel safer than risk.

Weaponization of digital archives

While dossierveillance often operates invisibly, its manifestations become painfully visible when digital archives – tweets, videos, emails – are weaponized to discredit, punish, or marginalize individuals. The Kevin Hart case demonstrates how digital archives function as living dossiers capable of re-entering public discourse years after their original production. The controversy was not driven by new conduct, but by the retrieval and reinterpretation of historical content under transformed cultural conditions.

The assumption that everything posted online is both permanent and retrievable creates an atmosphere in which the past is never safely behind us. Within this context, social media propaganda often exploits archival permanence, reframing old content to serve current ideological goals. The proliferation of cancel culture is both a symptom and a mechanism of this phenomenon. Individuals are “called out” based on past digital behaviors, which are often removed from their original context and reinterpreted under new moral or political frameworks (see Bouvier and Machin, 2021; Grunsven and Marin, 2024). While calls for accountability are not inherently unjustified, the process by which past content is uncovered, curated, and mobilized reflects dossierveillance in its affective and institutional forms. It also mirrors authoritarian dossier logic: the existence of a file, real or imagined, suffices to inhibit, judge, or penalize.

Consider the case of Kevin Hart, who in 2018 stepped down from hosting the Oscars after old homophobic tweets resurfaced from nearly a decade earlier. These tweets, which he had previously addressed and apologized for, were dredged up by online users, amplified through media outlets, and ultimately used to pressure the Academy to act (Dawson et al., 2023). The public controversy was not rooted in new behavior but in archival content – fragments of an older self-recontextualized in a different era of norms. This incident underscores how digital dossiers operate: not only can they be recalled and exposed, but they become malleable instruments of public judgment, shaped by whoever curates the narrative. While peer-initiated surveillance drove the initial exposure, it can be considered propaganda because the amplification of these archives serves to influence public perception and enforce social or moral norms. Furthermore, unlike physical records, digital archives are almost impossible to fully delete, making past content persistently retrievable and manipulable. This observation points not only to technological affordances of archival retrieval but also to broader cultural dynamics surrounding morality, reputation, and public shaming, suggesting that dossierveillance amplifies tendencies already present in contemporary media cultures rather than generating them ex nihilo.

In such contexts, propaganda does not take the form of overt lies or falsehoods. Instead, it manipulates the archival record to create moral clarity in a complex world. This weaponization reinforces social binaries – good versus bad, inclusive versus bigoted – flattening nuance and undermining the possibility of growth or transformation. It also fuels paranoia: if anything said online can be repurposed as future ammunition, the only safe course is hyper-vigilant curation of the self. For society, this produces a chilling effect not only on speech but on memory. People may begin deleting posts, avoiding political commentary, or distancing themselves from past beliefs – not because they have changed, but because the archive might be used against them. In this way, dossierveillance and propaganda combine to discipline not only what is said but what is remembered.

Algorithmic amplification and the rise of manufactured consensus

One of the most potent implications of the intersection between dossierveillance and social media propaganda is the way algorithmic systems amplify certain narratives, creating the illusion of consensus. The Cambridge Analytica scandal reveals how algorithmic amplification depends upon the continuous construction of behavioral dossiers. By harvesting personal data and categorizing users according to psychological profiles, propagandists were able to distribute emotionally calibrated messages tailored to specific anxieties and ideological predispositions.

Social media platforms are not neutral tools; they are designed to prioritize engagement, often by surfacing content that is emotionally charged, polarizing, or provocative. Propaganda takes advantage of this design by producing content tailored for virality – memes, hashtags, videos – that align with ideological goals while suppressing nuance and complexity. The dossiers compiled on users – through likes, clicks, search histories, and behavioral metadata – are essential to this process. These dossiers allow platforms and political actors to target users with finely tuned messages, reinforcing existing beliefs and deepening social divides. Over time, users are less likely to encounter opposing views, not because of deliberate censorship, but because the algorithm predicts they will not engage with them. This results in a form of epistemic closure, where people mistake engineered relevance for organic consensus. Above all, this closure is neither total nor irreversible: users may still encounter friction, contradiction, irony, and moments of reflexive doubt, underscoring that algorithmic influence operates probabilistically rather than deterministically.

A notable example of this dynamic is the 2016 Cambridge Analytica scandal. The British political consulting firm harvested the personal data of millions of Facebook users without their consent and used that information to build psychological profiles for micro-targeting during elections (Hinds et al., 2020). These profiles acted as dossiers, enabling propagandists to craft emotionally resonant messages aimed at manipulating voter behavior – particularly in the Brexit vote and the 2016 U.S. presidential election (Dmytro and Viacheslav, 2025; Kreiss and Barrett, 2020). Users were not aware of the dossiers being compiled on them, nor of the extent to which those dossiers shaped the content they consumed.

This example underscores the subtlety of modern propaganda: it does not necessarily require coercion or deception; it can operate through relevance, personalization, and affect. The messages individuals received were often emotionally manipulative, tailored to reinforce fear, anger, or ideological loyalty. Over time, the repetition of these narratives created a sense of collective agreement – what appears to be a democratic groundswell may, in fact, be a manufactured consensus. The long-term implication is a weakening of democratic discourse. If users inhabit separate realities, curated by algorithms and exploited by propagandists, meaningful debate becomes nearly impossible. The dossier becomes not only a record of past behavior but a tool for shaping future perception. It directs what one sees, believes, and shares –making the boundary between personal agency and manipulation increasingly difficult to discern.

As a result, democratic societies risk becoming panoptic echo chambers: citizens under the constant gaze of digital systems that not only record but also direct their thoughts and behaviors. The fusion of dossierveillance with social media propaganda does not merely threaten privacy; it alters the very conditions under which public opinion and political will are formed. However, these conditions remain historically contingent and politically contested, leaving open the possibility that regulatory reform, platform redesign, and collective resistance may yet reconfigure the balance between surveillance, propaganda, and democratic agency.