MALCOLM OGDEN
University of Richmond, USA

Abstract

This paper considers the stimulation afforded by ASMR NPC (autonomous sensory meridian response non-playable character) streams through the conceptual framework of ‘possession.’ First, it looks at several examples of popular ASMR NPC streams, noting the distinct manner in which ASMR NPC streamers become, like the non-player characters of role-playing video games, stuck in strange loops of repeated movements and phrases. I connect this framing of ASMR NPC streams/streamers, in its emphasis on an apparent objectification and loss of agency (or ‘possession’), to various popular commentaries on NPC streams, and video platforms in general. Drawing on media and communication studies scholarship on ‘digital possession,’ I then complicate this initial framing, asserting that ASMR NPC streams involve a more radically ambiguous and dispersed form of stimulation, in which the ‘possession’ of one individual or social group by another is subsumed by a more general sense of ‘possession,’ tied to the infrastructural character of audiovisual platforms under contemporary capitalism.

Keywords

TikTok, possession, interfaces, affect, free labor, immaterial labor

For the official version of record, see here:
Ogden, M. (2026). ASMR NPC: On Streaming Video and Digital Possession. Media Theory, 10(1), 101–124. https://doi.org/10.70064/mt.v10i1.1369

ASMR NPC:
On Streaming Video and Digital Possession

Introduction

In the summer of 2023, a spate of news articles attempted to make sense of the latest online trend: the “NPC stream.” An NPC stream, as these various articles explained, is a live interactive video feed (typically hosted on TikTok LIVE or an equivalent streaming service/platform) in which a sole content creator performs a limited number of predetermined, or preprogrammed, actions and/or phrases based on the monetized “Gifts” that individual viewers donate to the stream from one moment to the next. In the various articles on NPC streams that were published in 2023, and which gradually tapered off in 2024, commentators naturally focused on some of the more obvious distinguishing characteristics of NPC streams: the nonsensical procession of cartoonish and absurd catchphrases (e.g., TikTok creator Pinkydoll’s signature “ice cream so good!” “yes, yes, yes!” “gang gang,” etc.); the uncanny repeated movements and gestures; the often gendered and sexual undertones of the streams. Despite these common points of emphasis, many of these same articles also expressed some confusion as to what, exactly, NPC stream audiences “get” out of tuning into the streams, or why NPC streamers would be willing to seemingly objectify themselves in such a manner. And in the absence of any single satisfying answer, several of the articles seem to further ponder, in so many words: what does the phenomenon of NPC streams portend for modern society as a whole? As two commenters, quoted in a Forbes article on the phenomenon, lament: “Any time I accidentally happen upon a TikTok live, I feel like I am watching the world end,” and “this is like something you’d seen on a screen in the background of a verhoeven [sic] movie to show how dumb and insane the future is” (Di Placido, 2023).

In order to make sense of NPC streams, we might start by looking at similar categories of online videos. Fittingly, several popular articles on NPC streams note clear connections between NPC streams and ASMR—another prominent category of online videos and livestreams. ASMR, or “autonomous sensory meridian response,” is a term coined on an online health forum in 2010 that refers to a pleasant tingling sensation on the spine and scalp experienced by only some members of the population in response to particular auditory and visual stimuli (e.g., the sound of running water, chewing sounds, whispering, etc.). Since the term was first coined in 2010, it has become synonymous with a host of online audiovisual media forms and practices meant to provoke or “trigger” experiences of ASMR in viewer-listeners through the deployment of specific audiovisual stimuli (Gallagher, 2016). This has included, in more recent years, ASMR livestreams featuring “ASMRtists,” who generally employ a combination of roleplay scenarios and props for producing particular sounds (along with their own voices), along with some live audience interaction via text-based chat.

While there are a number of ways in which ASMR and NPC overlap, some of which I discuss at greater length in this paper, suffice it to say that since NPC streams went viral in summer 2023, a number of content creators who were already doing regular ASMR streams have even combined the two categories in doing “ASMR NPC” streams. In this paper, I will draw on existing scholarly work on ASMR, as well as scholarly work on affect, to help in analyzing the phenomenon of NPC streams, and the distinct form (or forms) of mediated and media-adjacent stimulation that NPC streams involve. Part of the initial argument laid out in the first section of this paper is that NPC streams and other similar categories of online videos such as ASMR derive at least some of their potency from the inherent ambiguity and instability that characterizes the relationality between the somatic (that is, the bodily or sensorial) and cultural-symbolic (i.e. meaning-related) features of audiovisual media, as experienced and transmitted via mundane media interfaces.

To address the characteristic dynamic of NPC streams wherein streamers are seemingly controlled from moment to moment by different audience members, however, this paper also engages extensively with scholarly work on possession. Spiritual or demonic possession refers to the religious belief found across a variety of religious traditions that the Devil and/or good and evil spirits not only exist in the world, but that they can invade human beings’ bodies, and in doing so, control human beings physically, and also influence their psychology and behavior (Levack, 2014: 1). The second section of this paper therefore features a reading of NPC streams as a form of possession. But controverting a number of overly reductive interpretations of NPC streams found in popular publications that mostly emphasize the objectified status of NPC streamers, the model of possession that this analysis employs is characterized by a notable degree of ambiguity, variously stemming from the role of media, affect, and collective cultural recognition in instances of possession. 

Furthermore, whereas certain religious framings of possession hinge on the idea that a demonic force can be expelled from the possessed, allowing for a return of the old “authentic” self, the theory of possession proposed in this article maps onto a set of conditions in which, I argue, there is no singular authentic self to be regained. Here, I argue that a media theory of digital possession that is sufficiently nuanced and embracing of ambiguity does not need to—and indeed, should not—ignore the dominant influence of Big Tech over contemporary social and political life, or the various escalating crises that have coincided with Big Tech’s rise to power. Nonetheless, neither audiovisual phenomena such as NPC streams, nor even the accumulated wealth and power of the platforms on which NPC streams are hosted, constitute a total and irreversible possession by demonic forces. Even while Big Tech has, in recent years, embraced its more “demonic”-seeming capacities for genocidal destruction (in Gaza and elsewhere), the mediated interactions seen in and around NPC streams, and numerous other categories of online videos, reveal a far more fluid and indeterminate set of relations than popular dismissals generally acknowledge.

I. Historical foundations: Media and possession

Popular articles on the phenomenon of NPC streams suggest that there is an appeal to the streams that operate below the level of cultural-symbolically inscribed meaning. One article, for instance, highlights the “nonsensical stream of words” and “robotic body movements” of NPC streams (Napolitano, 2023). Or to consider another example, a blog post included on the website for Qustodio (a software company whose products allow parents to surveil and control their children’s use of networked devices) notes: “it’s hard to tell what exactly [an NPC stream] even is, apart from watching a stranger online repeat the same gibberish over and over again [. . .]” (Lawrenson, 2023). A subheading for an article on NPC streams in The Guardian describes them as having “anarchic, erotic undertones” (Chan, 2023). The same article also describes the streams, in which content creators “rapidly string[ ] together nonsense catchphrases, over and over with the same cheerful expression and tone,” as “mesmerizing.” A writer for Vice uses the same term, “mesmerizing,” in their article on the phenomenon, which is largely a profile of streamer PinkyDoll (Cole, 2023).

In keeping with the previously mentioned user comments comparing watching NPC streams to watching the end of the world, these accounts of the ambiguously alluring quality of such content further suggest a certain trepidation regarding potent yet opaque and inaccessible forces—gleaned in and around particular audiovisual forms and categories, perhaps, but also video streaming platforms more generally. These hidden and only half-imagined forces, I argue, are best understood through the conceptual framework of possession. In social historian E. William Monter’s work on witchcraft in France and Switzerland during the Reformation, Monter refers to the period spanning the sixteenth and seventeenth centuries as constituting “the golden age of the demoniac” (Monter, 1976: 60). In The Devil Within: Possession and Exorcism in the Christian West, historian Brian Levack presents a history of possession in early modern Europe built on a similar understanding of the early modern period. One of the key arguments in Levack’s work is that printing, in particular, “played a crucial role in allowing the number of possessions to reach unprecedented proportions:”

The dissemination of demonological texts, sermons, possession narratives, bibles, reports of witchcraft prosecutions, and the printed versions of plays featuring demoniacs provided scripts for demoniacs to follow, precedents for those who diagnosed the symptoms of the possessed as demonic in origin, apocalyptical predictions to make the Devil’s fury seem plausible, and rituals for exorcists to use in their efforts to publicize the possession” (Levack, 2014: 258).

Levack’s specific framing of printed texts on possession as “scripts” in this description points to another key argument found in his work, which is that “all possessions, not just those that were feigned or otherwise volitional, were theatrical productions in which the demoniacs and also their families, neighbours, physicians, pastors, and exorcists played their assigned roles” (Levack, 2014: 30). Of course, Levack is not making the argument that possessions were literal theatrical productions, but his use of the theater as a metaphor for social performance nonetheless highlights the media-dependent nature of possession as a collectively enacted and socially constructed phenomenon.

In recent years, a growing body of scholarly work on spiritual possession and exorcism has responded to the continued prevalence of possession and exorcism across a range of contexts, indicating that perhaps the phenomenon of possession is not so far removed as some might assume. The so-called “secularist revisionists” working in the late twentieth-century observed that, contrary to the widely shared belief during the early modern period that religion would “disappear from both public and private social spheres,” religion had, in many instances, “never surrendered the private sphere and had even reappeared in the public sphere” (Possamai and Giordan, 2020: 2). More recently, a second wave of secularist revisionists have further developed this line of thinking, asserting that religion “had never really disappeared during modernity, even in the public sphere” (Possamai and Giordan, 2020: 2).

In their recent work on “digital exorcism,” Possamai and Gower (2024) emphasize the continued importance of media generally, and digital media specifically, to more recently observed increases in the prevalence of spiritual or demonic possession. Just as “the printing press, the radio, and the television have in the past provided a platform to expand one’s religious message,” they write, “the digital world is certainly not excluded and it today provides a way to upload ritual of exorcism” (Possamai and Gower, 2024: 10). Building on the insights of secularist revisionists, then, the use of possession as a conceptual framework in the analysis of distinctly non-religious phenomena (i.e. ASMR NPC streams) helps to highlight the regular circulation of religious practices and ideas, over time, into (supposedly) secular environments, and secular practices and ideas into religious ones, as an important dynamic to consider in the study of media more generally. One especially salient example of this sort of analysis can be found in media and communication scholar Stephan Gregory’s work on magical practices in 18^th century Bavaria.

Returning to the descriptions of NPC streams from the beginning of this section, it is noteworthy that two of those commentaries specifically describe NPC streams as “mesmerizing.” Mesmerism, historically, refers to the vitalist theory of “animal magnetism” developed by German physician Franz Mesmer in the 18^th century, which posits animal magnetism as being “an occult force or invisible fluid emanating from his body” that “permeated the universe, deriving especially from the stars” (Encyclopedia Britannica, 2005). In his work, Stephan Gregory provides a media theoretical explanation for why and how “the popular exorcist practices performed by the catholic priest Johann Joseph Gassner were eventually discredited, outlawed, and superseded by the enlightened, ‘scientific’ system of Franz Anton Mesmer’s “Animal Magnetism” (Gregory, 2015: 1). Gassner’s exorcisms were, as Gregory writes, “something of a fairground attraction,” and differed in certain respects from other approaches to exorcism of the time through their focus on the physical and corporeal:

Because it is not simply about the transmission of information, but about the transport of affects, Gassner’s procedure requires the production of an effective contact, a physical act of touching. It presupposes a bridge or a channel to be created, so that the demonic energy can flow off. It requires an intermediate thing, a medium that sets up the connection and controls the exchange (Gregory, 2015: 14).

In contrast with other approaches to exorcism from the same time period that focused on the specific meanings conveyed by the possessed through their spoken words, Mesmer’s treatments were similar to Gassner’s in that they placed a stronger emphasis on “corporeal touch, the physical connection between the patient and the magnetizer” (Gregory, 2015: 20). Hence, when a prohibition on exorcism was decreed for all of Bavaria in 1775, Mesmerism easily “filled in” the void, serving a similar function as Gassner’s exorcisms had previously but without any overtly religious features.

As Gregory acknowledges, the basic distinction between “two ways of understanding supernatural communication” that his analysis outlines can be likened to a number of well-worn conceptual distinctions that, he writes, “still govern today’s media and communication studies: medium vs. message, channel vs. content, hardware vs. software […]” (1). Over the past several decades, scholarly work on affect has repeatedly invoked such distinctions, while also interrogating the notion that one side ever exists, or can be meaningfully grasped, in isolation from the other. For example, in Brian Massumi’s (2002) foundational writing on the “autonomy of affect,” Massumi distinguishes between “intensity” and “qualification” as two orders of connection “operating in parallel” (24). “Intensity,” he writes, is “embodied in purely autonomic reactions most directly manifested in the skin at the surface of the body, at its interface with things.” Qualification, meanwhile, has to do with “signification and coding” (25), and, according to Massumi, should not be dismissed outright as being merely “false or unreal,”  but its “sphere of applicability” should nonetheless be understood as “limited to a particular mode of existence, or a particular dimension of the real (the degree to which things coincide with their own arrest” (7).

Affect theory’s influence on and overlapping with feminist new materialisms, critical posthumanism, and speculative realism, stems in part, then, from this basic distinction, in speculatively mapping a world of nonhuman (or “more-than-human”) forces and relations not entirely beyond, but also not clearly or easily reducible to, human scales of meaning. Sara Ahmed (2010), for instance, in her writing on “happy objects,” describes the process through which objects “circulate as social goods” in specific cultural contexts, and “accumulate positive affective value” in the process (30). Key to Ahmed’s distinct formulation of affect is the radical contingency of affective encounters—“the messiness of the experiential, the unfolding of bodies into worlds, and the drama of contingency, how we are touched by what we are near” (30)—combined with the recognition that “judgement about certain objects as being ‘happy’” has, in many instances, already occurred prior to one’s seemingly serendipitous encounters with them. The messy and often conflict-riven process by which affective encounters are translated into words, symbols, and narratives is thus never truly “finished,” but rather it continually and recursively unfolds throughout one’s everyday embodied existence.

Part of my specific interest in the intersection of media studies, affect theory, and emergent forms of digital possession, is methodological. In scholarly work on YouTube and TikTok, a notable divide has to do with whether, how, and why one should bother paying attention to the specific “content” of particular pieces or categories of audiovisual media online, or whether one should instead prioritize analyzing more generally applicable aspects of platforms that might also be less apparent to platform users: the underlying or “macro-level” social, political, and economic conditions out of which platforms emerged, historically, and to which they have further contributed. As Alison Hearn and Sarah Banet-Weiser (2020) note, “much recent media theory has moved away from discussions of aesthetics, narrative, and representation, arguing that the primary function of digital technologies is now organizational, allocative, infrastructural, or logistical” (2).

A number of scholars have, in recent years, reasserted the continued importance of aesthetics and representational analysis to the study of platforms—in some instances, not despite but in light of their more “infrastructural” and “organizational” aspects. For instance, in The Influencer Factory: A Marxist Theory of Corporate Personhood on YouTube, Bollmer and Guinness (2024) organize their Marxist analysis of specific influencers, and of “influencer” culture more generally, according to some of various different “backgrounds” that “constantly appear in the videos of influencers,” such as “the home,” “the car,” “the warehouse” and “the ‘market’” (43). By foregrounding these “backgrounds,” Bollmer and Guinness frame their analyses of specific influencers, such as Emma Chamberlain, Jeffree Starr, and Mr. Beast, as constituting “an entry point into a set of relations that far exceed social media,” and as, more precisely, “a route into the concrete materiality of contemporary capitalism” (7).

Media scholar Susanna Paasonen’s (2010; 2021; 2023) work has especially informed the model of affect employed throughout the analysis portion of this paper. In her work on video pornography, Paasonen (2010) utilizes affect to consider the complex interrelations of video, text, meaning, and embodied sensation found in and around pornographic media. Affect, for Paasonen, helps to account for “the simultaneous “shapelessness” and sharpness of sensation” (26) in analyzing such phenomena, which she describes as “rife with inner contradictions and tensions” (256). In more recent work, Paasonen (2021) has further stressed the importance of “foregrounding ambiguity and contextual nuance over generalizations and reductive conclusions” (4) in the study of digital media. Echoing Hearn and Banet-Weiser, Paasonen notes that while certainly “networked media reproduces, feeds, and accelerates capitalist accumulation within the data economy,” this now well-noted dimension of networked media “does not exhaust its meanings, functions, or political potentialities” (8-9). Noting the recent uptick in the commercial development and use of sentiment analysis technologies, Paasonen (2023) critiques the manner in which affect has become “positioned as guarantor of authenticity or even as a subjective truth concerning objects, situations, and settings—and is monetized as such” (85).

Crucially, employing a model of affect as ambiguity does not mean lapsing into blurry indistinction, or somehow escaping from meaning entirely. For Paasonen, rather, it means allowing the micro- (“the quotidian, the messy, and the felt”) and macro-levels (e.g., “the dynamics of data capitalism, the algorithmic polarization of sociability, and the political economies of targeted advertising”) of analysis to meet (100), without assuming an overly simplistic relationality between the two from the outset.

Following Bollmer and Guinness, Hearn and Banet-Weiser, Paasonen, and others, my own analysis of NPC streams will proceed from the basic assertion that, in studying online videos and video platforms, paying attention to the specific sounds and images coming through speakers and screens matters. Closely analyzing particular pieces or categories of audiovisual media, as well as the textual discourses and affective attachments that surround them, can help us to develop more nuanced understandings of the specific systems of relations that platforms involve, which in some instances extend to broader social and political structures, while in other instances their status relative to said phenomena is less clearly defined.

II. Video streaming as possession

An initial description of some of the main components of an NPC stream, presented here, will serve as a useful point of reference in the analysis that follows. The typical NPC streamer streams from their phone. Most stream viewers watch the streams from their phones or other mobiles devices. NPC streams therefore typically feature a vertical aspect ratio, meaning that the video is taller than it is wide. Within this video frame, which is just one component of several that make up a video stream, the streamer appears and performs for their audience. While preparing this article, one specific stream that I watched live was by streamer Matthew Zimmerman, whose channel name is JustAnotherSkywalker. As his channel name would suggest, during his streams, Zimmerman usually dresses up as Anakin Skywalker, a character from the Star Wars franchise, and many of the specific phrases and gestures that he performs are taken from the Star Wars movies (e.g., “I don’t like sand—is that corny?” in response to the “It’s corn” Gift). As viewers donate monetized Gifts from moment to moment, the Gifts appear as animations overlaid over the video feed of streamers’ embodied performances.

The specific stream that I watched lasted for over an hour, though Zimmerman “performed” as an NPC for maybe only half of the stream’s duration; during the latter half, he conversed with viewers in a relaxed and informal manner. Since lower-value Gifts are more commonly given, much of the NPC segments of the stream that I watched consisted of a handful of simpler, less energy-intensive actions/phrases being repeated over and over. But a variety of more elaborate performances corresponding to higher value “gifts” occurred with surprising regularity. These included elaborately choreographed demonstrations of Zimmerman’s skills with light-up model lightsabers that emitted sound effects when swung about, but also hip-hop dance routines with no discernible connection to Star Wars beyond Zimmerman’s attire.

To the right of the video screen at the top of the page, the “Top Viewers” board displays the usernames of the five or so “top viewers” listed in descending order alongside a calculated numerical score for each. Per TikTok’s FAQ, top viewers are ranked “based on Gifts count and watch time.” If a viewer has not donated any Gifts, they will still be ranked based on “how long they have been watching the LIVE.”  Beneath the “Top Viewers” section, the live chatbox appears, featuring a chaotic flow of viewer comments. While Zimmerman remained mostly narrowly focused on responding to Gifts during the NPC portion of his stream, he also responded to specific comments in the chat several times throughout. Looking through the various comments from the Andrew Zimmerman stream that I watched, one gets the sense that many of the commenters themselves are, in some sense, performing for each other, for the streamer, and/or for the many viewers who tune in to NPC streams but do not comment or donate. One user comment reads: “Imagine walking by and enjoying your day and you see this guy.” Another user makes a request: “Can you say happy birthday Peyton.” Another guesses Zimmerman’s location based on his surroundings: “This has to be the Midwest. Ohio or PA.”

The last feature of NPC streams that should be mentioned here is the red heart-shaped “Like” button, situated beneath the live chatbox. While the “Like” button on TikTok livestreams functions similarly to those found across most social media platforms, one crucial difference is that there is no limit to the number of times that an individual viewer can press the like button. Rather, pressing the button as many times as possible is explicitly encouraged by streamers, as the total number of likes that a stream has received over the course of the stream appears as a constantly updating number in the top lefthand corner of the stream. Similar to “number of viewers” or “number of subscribers” on different social media platforms, the total number of likes on a video livestream on YouTube or TikTok is a crucial metric by which the popularity and success of a stream is measured—and by extension, algorithmically recommended (or not recommended) by the platform to potential viewers.

As the above description illustrates, NPC streams, not unlike the various forms of historical and contemporary possession discussed in the previous section, are a thoroughly mediated and media-dependent phenomenon. Media have provided a range of cultural meanings with which to make sense of NPC streams, but, beyond the level of cultural-symbolic meaning, media and media interfaces also serve as constitutive elements with respect to the physical, affective, ambiguous, and at times unconscious relations that NPC streams tend to involve. Digital possession attunes us to both realms of analysis as inextricably bound up with each other. As Susanna Paasonen (2023) writes, “[a]nalytical zooming in—solely—on either the micro or the macro will remain partial in its grasp, for it is precisely the dynamic, contradictory and ambiguous entanglements of the structural and the individual, the societal and the personal, the corporate and the intensely felt that make both data culture and the self” (100).  Such an approach does not mean “doing away with the mess” but rather “diving into it” (100).

In diving deeper into the mess of NPC streams, we might consider some of the more “obvious” cultural-symbolic meanings and associations at play, starting with the term “NPC” itself. The “NPC” in “NPC streams” is a reference to “non-playable characters” in video games like Oblivion and Grand Theft Auto, and the seemingly mindless manner in which NPCs will repeat certain actions or certain lines of dialogue. By definition, players typically do not have immediate or direct control over NPCs. Instead, players interact with NPCs, though the repetitive and formulaic character of these interactions can make such a distinction seem relatively arbitrary. NPCs are, by and large, without any discernible “agency” or “autonomy,” as we might conventionally understand the term. In any game with a relatively prescriptive set of goals or objectives, NPCs can be considered as more or less equivalent to an in-game item that one might utilize in the pursuit of achieving their diegetically defined goals.

But of course, NPC streamers are no more separate from the “real world” than even the fictional characters of videogames—a truth that is likely apparent to the viewers who interact with NPC streamers either during the NPC portions of their streams or when the streamer is no longer “in character.” As the comments on Zimmerman’s stream quoted above demonstrate, the chosen location for a stream is just one of the ways that viewers relate to streamers. Similarly, the NPC performance of NPC streams often takes place in public spaces that viewers might recognize, or which might provoke interactions with passersby. Which is to say that the strictly “in-game” model of an NPC elaborated above clearly does not account for the variety of forces at play—though it does perhaps, as a metaphor, explain some of the more extreme reactions to NPC streams.

In light of the sheer variety of symptoms that texts from early modern Europe have attributed to possession, Levack (2014: 15-16) identifies two “essential features” of possession that characterize all of the case studies included in his analysis, which are: (1) abnormality (which he further subdivides into the “psychosomatic” and “psychopathological”), and (2) the conscious and explicit identification of possession as an explanation of one person’s behavior by at least one other person. In applying this framework to NPC streams, specifically, I will start with the latter of the two “essential features” of possession indicated by Levack, having to do with recognition (as closely related to “identification” and/or “explanation”) as a shared and collective social and cultural process. The stimulation of possession, understood in this register, involves the playful exchange of symbols and meanings in and around the words, gestures, and behaviors of NPC streamers from one moment to the next.

Returning briefly to the name “NPC,” it should be noted that NPC streams represent merely one more recent variety in a range of “NPC” video content that can be found online, which generally involve the comedic and/or highly choreographed performance of different gestures and interactions à la video game NPC (Hill, 2023). Hence, one popular news article describes NPC streams as “kind of an in-joke” (Chan, 2023). Being “in on the joke,” and thereby participating in the “recognition” of digital possession, requires, in this instance, having some familiarity with video games (or, at the very least, a certain online offshoot of video game culture). In recent years, “NPC” has also taken on another meaning in online political discourse—that is, someone who holds generic beliefs and opinions on a subject. In online political discourse, the term has at times served as an expedient means for right-wing commentators to criticize “political correctness” or “wokeness” (Know Your Meme, 2023).

In both of the above definitions, NPC serves as a kind of in-group/out-group signifier—in one instance, for gamers, and for another, political ideology. In either case, what is important is that the naming and categorizing of NPC streams, like the term NPC itself, is, and has been, a collective cultural endeavor that is very much part of the “possession” at hand. But there are, of course, other varieties of collective cultural recognition at work in and around NPC streams beyond those explicitly attributed to the term “NPC”. One especially dominant strain of commentary found in the various articles on NPC streams published in 2023-2024 focuses explicitly on gender, sexuality, and emergent forms of affective and emotional labor online. For instance, a CBS News article published in July 2023 describes “NPC Content” as “a niche category that incorporates elements of ASMR, and that some say contains sexual undertones” (Napolitano, 2023). As several articles from around the same time period point out, both PinkyDoll and Cherry Crush (often hailed as “top practitioners” of the genre) did have OnlyFans accounts at the time, but given that any explicit sexual content on TikTok would be quickly removed, “the two spaces [were] kept entirely separate.”

In both popular and scholarly writing on ASMR, it has often been noted that many of the most popular creators in the genre are, or have been, “conventionally attractive” women with relatively lighter skin complexions. The use of roleplay scenarios by ASMRtists, which commonly involve forms of synthetic intimacy, further emphasizes the constitutive role that racialized and gendered forms of labor have played in shaping ASMR over the past 15 years. As media scholar Emma Waldron writes,

“[…] the most common settings for ASMR role-play videos are medical environments, salons and spas, and retail or other customer service settings. Women are also more likely than men to produce role-play and personal attention videos. Thus, although not all ASMR videos involve intimate narrative settings, and although not all ASMRtists are women, the digital spaces of care in ASMR YouTube videos are distinctly gendered, in that they are more likely to be populated by women” (Waldron, 2016).

Of course, NPC streams, like ASMR videos, are not exclusively produced by women, nor do they always (or even frequently) include an overtly sexual dynamic. My intention in raising this particular set of topics is not to conclusively pinpoint how exactly NPC streams relate to discourses pertaining to gender, sexuality, or emergent forms of labor in any absolute or definitive manner. But what these topics do help to illustrate is the role that discursive norms play in digital possession, by providing some schema against which words, behaviors, and bodily movements or appearances can be normatively evaluated. Hence, the first “essential feature” of possession identified by Levack (2014: 15) has to do with “abnormality.”

The physiological symptoms of historical instances of demonic possession, such as “convulsions, contortions, paralysis, loss of vision, and levitation,” were, according to Levack, “pathological in the sense that they deviated from established medical norms.” A perceived or recognized abnormality “characterized the verbal and behavioural signs of possession” as well, which included “speaking in previously unknown foreign languages, clairvoyance, and the public violation of conventional standards of conduct.” Levack’s including “the public violation of conventional standards of conduct” in this list especially stands out in light of the fact that when various women claimed to have achieved “mystical union with the divine” during the early modern period, he explains, “the reclassification of divine possessions as demonic became almost routine by the early seventeenth century” (Levack, 2014: 258). This reclassification was fueled by not just the “similarities between the symptoms of divine possession that these women cultivated and those attributed to the Devil,” but also “the misogynistic assumption that women were more likely than men to have commerce with demons.”

As it is informed by scholarly research on possession as a historical phenomenon, a media theory of digital possession helps to map how certain social groups (e.g., women, children, sex workers, the disabled) are or have been discursively framed as uniquely vulnerable to certain forms of “harm,” and how this framing fuels further paternalistic policing of the attitudes and behaviors of people from marginalized groups. An article entitled “Leave sex workers out of your TikTok discourse” published in TechCrunch around the time that NPC streams went viral laments how “[e]very time a woman monetizes the absurd, she’s accused of making fetish content” (Sung, 2023). This trend, the article continues, is furthermore “a symptom of the anti-sex moral panic sweeping pop culture—and this time it’s coming for TikTok’s NPC creators.” The real cause behind some of the most prominent criticisms of the genre, the article explains, is “the fact that these creators—most of whom are attractive young women—manage to profit off of being weird.” As the same TechCrunch article explains, criticisms of NPC streamers like PinkyDoll seem to tie in with a range of recent cultural anxieties surrounding networked media, gender, sexuality, and labor:

“It’s an NPC control kink,” one Twitter user commented on a video of Sinon’s stream. “Don’t look [too] much into it … I wish I didn’t. Butttt basically they’re controlling her like a video game character.”

Others accused her of “catering to pedos” because she initially told viewers that she’s 19 years old, not 27. One commentary creator described NPC TikTok as “just another gateway drug for getting boys hooked on fetish porn.” Another Twitter user called Sinon an “NPC sex doll for incels” (Sung, 2023).

There are no easy resolutions to these discourses, which, in their ambiguous affective dimensions, only intensify the further circulation of constellated text, sounds, and images on platforms like YouTube, TikTok, and Twitch. After all, as Susanna Paasonen (2023) writes, “affective complexity is precisely key to the attraction that much social media holds as users take pleasure in witnessing other people’s opinions, or reading narratives of grief, devastation and suffering” (93).

In conjunction with more disapproving interpretations of NPC streams, there is one clear sense in which NPC streamers have been framed, in many of the popular articles discussing the phenomenon, as laudable, if not exactly “normal.” As a Vice article explains,

[PinkyDoll] told me she spends six hours a day on TikTok, seven days a week. “It’s not exhausting,” she said. “I kinda love to go live and reacting to gifts. There are so many and the views going up is boosting me.” She hit one million views on one live streaming session, she said. “It’s keeping me energetic.” When she first started, she made $250 a day. But when she went viral this month, that turned into $7,000 a day and more (Cole, 2023).

Here, PinkyDoll is presented as exceptional for their success in achieving, through this emergent form of mediated labor, both high visibility and high income. Hardt and Negri (2001) define immaterial labor as ‘labor that produces an immaterial good, such as a service, a cultural product, knowledge or communication’ (Hardt and Negri, 2001: 290). As Maurizio Lazzarato (2006) explains, immaterial labor does not exclusively refer to a specific social class or occupation, but rather it encompasses a range of more generally dispersed changes in postindustrial societies wherein the increasing prevalence of networked and computational systems coincides with a gradual blurring of the distinctions between work and leisure, public and private. Under such conditions, as Paulo Virno writes, “[e]xcess time—that is, a potential wealth—manifests itself as misery: forced redundancy, early retirement, structural unemployment (triggered by investments, not the lack of them) and the proliferation of hierarchies” (Virno, 2007: 5).

It is perhaps unsurprising, then, that Pinkydoll, as the Vice article explains, “started using TikTok to make money” after returning to Montréal after her stepfather’s funeral and finding that she had lost the cleaning company that she had previously run (Cole, 2023). She explains, “I needed money to feed my kid and pay the bills. I had no job. I decided to put all my effort on TikTok to make money and I wasn’t expecting to go viral and make all that money. It was a shock.” In her earlier writing on free labor, Tiziana Terranova (2000) highlights the increasingly common trend under post-Fordism of individuals voluntarily performing tasks on the internet without any monetary compensation, which large companies ultimately benefit from. In contexts where post-Fordism has already been in place for several decades, she explains, this tendency results from the interests of children and young people being consistently steered toward business interests from birth. While Terranova’s writing on free labor preceded the rise of the increasingly centralized internet of Web 2.0, it remains an important theoretical framework to consider in the present in that, as Dyer-Witheford and Mularoni write,

Data extraction can only occur because users are not passive, but active: logging on, posting, playing, socializing and creating. The conversion of this activity into capitalist value—monetization—certainly demands complex, layered mediations, through the paid work of technicians and moderators, and sale and purchase of online advertising. Yet the recent development in partially paid and precarious networked labour shows how thin the line separating users from workers has become […] (Dyer-Witheford and Mularoni, 2025: 35).

While PinkyDoll and other popular NPC streamers are celebrated (at least by some) for their having achieved at least temporary reprieve from conditions of generalized precarity, they nonetheless harken us to consider yet another crucial aspect of digital possession, which is the role of social crisis. Drawing on Michel de Certeau’s study of the mass exorcism of Loudun in seventeenth-century France, which “links the increase and spread of exorcism with periods of profound social change,” Possamai and Giordain (2020: 6) assert that “the study of exorcism today offers an opportunity to describe not only the crisis and the drama of a single individual, but also the crisis of the society in which the possessed lives.” To conclude this paper, then, I will consider some of these crises, and the possibilities for exorcism that they might portend.

III. Conclusion: Possession reconsidered

Digital possession originates from conditions in which opaque and only semi-graspable processes of determination abound, as well as speculative yet intensely felt fears regarding those partially imagined processes. In his seminal writings on “determination,” cultural studies theorist Raymond Williams (1977) provides an especially cogent explanation of the concept:

For in practice determination is never only the setting of limits; it is also the exertion of pressures. As it happens this is also a sense of ‘determine’ in English: to determine or be determined to do something is an act of will and purpose. In a whole social process, these positive determinations […] are by no means only pressures against the limits, though these are crucially important. […] They are also, and vitally, pressures exerted by new formations, with their as yet unrealized intentions and demands (87).

Here, Williams’s description of determination resembles neo-Spinozan conceptualizations of affect as potentiality, with the exercise of power functioning not solely through an external delimiting of possibilities for action, but also through enabling and encouraging other particular pathways. The end-result of such processes, as they unfold at various scales, is by no means a foregone conclusion, as “new formations” arise with “yet unrealized intentions and demands.” In an interview published in Politics of Affect, Brian Massumi (2015) provides a similar explanation of how power functions in light of the transformations that have taken place in relatively affluent capitalist societies since World War II. Citing the works of Antonio Negri, in particular, he explains:

The argument is that capitalist powers have pretty much abandoned control in the sense of ‘power over’. […] It starts working directly on bodies’ movements and momentum, producing momentums, the more varied and even erratic the better […] (20).

Massumi’s account (paraphrasing Negri), considered in tandem with Williams’s observation that both forms of determination are articulated with the same subjective framework of the “individual” speaks to, among other things, today’s influencer culture. In the world of influencers, notions such as “individuality” and “authenticity” serve as the justification for influencers’ qualifications as intermediaries between large companies engaged in the selling of commodities and influencers’ audiences, who are interpellated as potential consumers (Bollmer and Guinness, 2019).

In the case of NPC streams, similarly, a more absolute and unidirectional framing of possession would tell us that within the figure of the NPC streamer there perhaps still exists a discrete and separate self, but that streamers’ outward actions, including both words and gestures, can be attributed to a more profound yet concealed loss of agency. Per this naïve reading of NPC streams as possession, an exorcism of the external agent would allow for the restoration of the true or authentic self. Returning once more to Williams’s writing on determination, however, we might approach such a framing of possession with some skepticism.  He writes:

Society is then never only the ‘dead husk’ which limits social and individual fulfillment. […] Any abstraction of determinism, based on the isolation of autonomous categories, which are seen as controlling or which can be used for prediction, is then a mystification of the specific and always related determinants which are the real social process—an active and conscious as well as, by default, a passive and objectified historical experience (87).

In their writing on digital possession, Possami and Gower note that the COVID-19 pandemic represents just one crisis of several that has “without a doubt accelerated this shift towards the digital sphere” (Possamai and Gower, 2020: 2). In a recent article, similarly, Franco Berardi (2023) decries the “condition of isolation” that “the pandemic has normalized and almost institutionalized,” but which “neoliberal and digital acceleration,” he argues, had already brought about to a considerable degree even prior to the pandemic. Hence, “desire in our age” is, according to Berardi, “defined” by not only “neoliberal and digital acceleration,” but the rise of “contemporary identitarian reactionary movements” as well. 

Amidst such developments, digital possession points to, among other things, the likely function of myth and metaphor in making sense of the myriad social, political, and technological processes that are perpetually shaping our world at scales and levels beyond what individual human beings can comprehend or perceive. AI scholar Beth Singler (2020) has looked at the “theistic conceptions of AI” undergirding users’ perceptions of their own navigation of algorithmic systems, which are well encapsulated by the popular expression “blessed by the algorithm.” Scholarly work on “algorithmic imaginaries” (Bucher, 2016) and “algorithmic gossip” (Bishop, 2019) similarly emphasizes the opaque nature of algorithmic recommender systems, but also the nigh inescapable compulsion to try to anticipate our being manipulated by recommender systems toward ends that do not serve us.

What seems apparent is that we cannot rely on false notions of a pure and authentic self that that will somehow be freed from any and all oppressive influences via some sort of exorcism—at least in how it is commonly understood. Nonetheless, what a theory of digital possession can also help us to grasp is that whatever possibilities are in fact available to us can and will be found in the particularities of the present—wherein, like in a video livestream, we can stop repeating the same old words and gestures and perhaps grasp the potential for collective mobilization manifested in each passing moment. In their work on religious rituals of trance possession in Cuban Lucumí/Santería, and Brazilian Candomblé, Roberto Strongman (2019: 2-3) uses the concept of “transcorporeality” to describe “the distinctly Afro-diasporic cultural representation of the human psyche as multiple, removable, and external to the body that functions as its receptacle.” Bearing this alternative conceptualization of possession in mind, a media theory of possession lastly invites us to explore, in action, and beyond the sharing of myths, what else a body can do—and what pieces of ourselves we could perhaps live without.

References

Ahmed, S. (2010) ‘Happy Objects’, in M. Gregg and G. Seigworth (eds.), The Affect Theory Reader. Durham, NC: Duke University Press, pp. 29-51.

Berardi, F. (2009) The Soul at Work: From Alienation to Autonomy. Cambridge: MIT Press.

Berardi, F. (2017) Futurability: The Age of Impotence and the Horizon of Possibility. Brooklyn: Verso.

Berardi, F. (2023) ‘Hyper-Semiotization and De-Sexualation of Desire’, e-flux issue #133. Available at: https://www.e-flux.com/journal/133/514287/hyper-sem‌io‌t‌‌ization-‌and-de-sexualization-of-desire-on-felix-guattari/ (Accessed: May 15 2025).

Bishop, S. (2019) ‘Managing visibility on YouTube through algorithmic gossip’, New Media & ‌Society, 21(11-12), pp. 2589-2606. https://‌doi.org/‌‌10.1177/1461444819854731

Bollmer, G. and Guinness, K. (2024) The Influencer Factory: A Marxist Theory of Corporate Personhood. Stanford, CA: Stanford University Press.

Bollmer, G. (2019) Materialist Media Theory: An Introduction. London: Bloomsbury.

Chan, W. (2023) ‘$7000 a day for five catchphrases: the TikTokers pretending to be “non-playable characters’”, The Guardian, July 19. Available at: https://www.theguardian.com/technology/2023/jul/19/tiktok-npc-streaming-live-stream-pinkydoll (Accessed: May 15 2025).

Cole, S. (2023) ‘Viral TikTok NPC Streamer PinkyDoll Doesn’t Care What You Think’, Vice, July 17. Available at: https://www.vice.com/en/article/viral-tiktok-npc-streamer-pinkydoll-doesnt-care-what-you-think/ (Accessed: May 15 2025).

Di Placido, D. (2023) ‘TikTok’s NPC “Ice Cream So Good” Meme, Explained’, Forbes, July 20. Available at: https://www.forbes.com/sites/danidiplacido/2023/07/20/tiktoks-npc-ice-cream-so-good-meme-explained/ (Accessed: 15 May 2025).

Dyer-Witheford, N. and Mularoni, A. (2025) Cybernetic Circulation Complex: Big Tech and Planetary Crisis. Brooklyn: Verso.

Encyclopedia Britannica (2005) ‘Animal magnetism’. Available at: https://www.britannica.com/science/animal-magnetism (Accessed: 15 May 2025).

Gallagher, R. (2016) ‘Eliciting Euphoria Online: The Aesthetics of “ASMR” Video Culture’, Film Criticism 20(2), pp. 1-15.

Glaze, V. (2023) ‘How much does PinkyDoll make on TikTok? NPC TikToker and OnlyFans star reveals earnings’, Dexerto. Available at: https://www.dexerto.com/entertainment/npc-tiktoker-onlyfans-star-pinkydoll-reveals-staggering-earnings-2218222/ (Accessed: 15 May 2025).

Gregory, S. (2015) ‘Media in Action: From Exorcism to Mesmerism’, communication +1 4(1), pp. 1-30.

Guranathan, J. (2023) ‘Jessie Gurunathan: Tik Tok star Pinkydoll made me feel like my Boomer parents’, The New Zealand Herald, August 5. Available at: https://www.nzherald.co.nz/nz/jessie-gurunathan/YFGFWLCM3RH4PJMDIYESGGYNJQ/ (Accessed: 15 May 2025).

Hardt, M. and Negri, A. (2001) Empire. Cambridge: Harvard University Press.

Hearn, A., & Banet-Weiser, S. (2020) ‘The Beguiling: Glamour in/as Platformed Cultural Production’, Social Media + Society, 6(1). https://doi.org/10.1177/2056305119898779 (Original work published 2020)

Hill, E. J. (2023) ‘People are pretending to be ‘NPCs’ on TikTok and it’s not just weird, it’s also lucrative’, The Conversation, August 8. Available at: https://theconversation.com/people-are-pretending-to-be-npcs-on-tiktok-and-its-not-just-weird-its-also-lucrative-210795

Know Your Meme (2023) NPC Streaming. Available at: https://knowyourmeme.com/memes/npc-streaming (Accessed: May 15 2025).

Lawrenson, E. (2023) ‘What is NPC streaming? Why is it so popular?’ Qustodio by Qoria, August 29. Available at: https://www.qustodio.com/en/blog/what-is-npc-streaming/ (Accessed: May 15 2025)

Lazzarato, M. (2006) ‘Immaterial Labour’, in Virno, P. and Hardt, M. (eds.) Radical Thought in Italy: A Potential Politics. Minneapolis: University of Minnesota Press, pp. 133-150.

Lazzarato, M. (2014) Signs and Machines: Capitalism and the Production of Subjectivity. Cambridge: Semiotext(e).

Levack, B. (2014) The Devil Within: Possession and Exorcism in the Christian West. New Haven: Yale University Press.

Massumi, B. (2002) Parables for the Virtual: Movement, Affect, Sensation. Durham, NC: Duke University Press.

Massumi, B. (2015) Politics of Affect. Hoboken, NJ: Wiley.

Monter, E.W. (1976) Witchcraft in France and Switzerland: The Borderlands during the Reformation. Ithaca, NY: Cornell University Press.

Napolitano, E. (2023) ‘In latest TikTok fad, creators make big bucks off NPC streaming’, CBS News, July 27. Available at: https://www.cbsnews.com/news/tiktok-what-are-npc-videos-pinkydoll-cherry-crush/ (Accessed: 15 May 2025).

Park, G. (2023). ‘Why TikTokers are pretending to be robots and saying “ice cream so good”‘, The Washington Post, July 17. Available at: https://www.washingtonpost.com/arts-entertainment/2023/07/17/npc-pinkydoll-tiktok-live-streaming-robot/ (Accessed: 15 May 2025).

Paasonen, S. (2011) Carnal Resonance: Affect and Online Pornography. Cambridge: MIT Press.

Paasonen, S. (2022) Dependent, Distracted, Bored: Affective Formations in Networked Media. Cambridge, MA: MIT Press.

Paasonen, S. (2023) ‘Ambiguous Affect: Excitements That Make the Self’, in Seigworth, G. and Pedwell, C. (eds.) The Affect Theory Reader 2: Worldings, Tensions, Futures. Durham, NC: Duke University Press.

Possamai, A. and Giordan, G. (2020) ‘Introduction’, in A. Possamai and G. Giordan (eds.) The Social Scientific Study of Exorcism in Christianity. Pop Culture, Religion and Society Book Series. London: Springer, pp. 1-16.

Possamai, A., and Gower, R. (2024) ‘Digital exorcism and morality’, Journal of Sociology 0(0), pp. 1-13.

Seigworth, G. and Pedwell, C. (2023) ‘Introduction: A Shimmer of Inventories’, in Seigworth, G. and Pedwell, C. (eds.), The Affect Theory Reader 2: Worldings, Tensions, Futures. Durham, NC: Duke University Press.

Singler, B. (2020) ‘“Blessed by the algorithm”: Theistic conceptions of artificial intelligence in online discourse’, AI & Society 35(4), pp. 945-955.

Stokel-Walker, C. (2023) ‘The “NPC” livestream TikTok trend helping creators earn cash’, BBC, August 14. Available at: https://www.bbc.com/worklife/article/20230811-the-npc-livestream-tiktok-trend-helping-creators-earn-cash (Accessed 15 May 2025).

Strongman, R. (2019) Queering Black Atlantic Religions: Transcorporeality in Candomblé, Santería, and Vodou. Durham: Duke University Press.

Sung, M. (2023) ‘Leave sex workers out of your NPC TikTok discourse’, TechCrunch, July 21. Available at: https://techcrunch.com/2023/07/21/pinkydoll-npc-tiktok-sex-work-not-fetish-content/ (Accessed 15 May 2025).

Terranova, T. (2000) ‘Free Labor: Producing Culture for the Digital Economy’, Social Text 18(2), pp. 33–58.

Terranova, T. (2022) After the Internet: Digital Networks between Capital and the Common. Boston: MIT Press.

Virno, P. (2007) ‘General Intellect’, Historical Materialism 15(1), pp. 3-8.

Waldron, E.L. (2016) ‘“This FEELS SO REAL!” Sense and sexuality in ASMR videos’, First Monday 22(1).

Williams, R. (1977) Marxism and Literature. Oxford, UK: Oxford University Press.

Young, F. (2016) A History of Exorcism in Catholic Christianity. London: Springer.

Malcolm Ogden is a media theorist whose work explores the interrelations of digital media, bodily sensation, and changing forms of work and leisure under contemporary capitalism. His work has been published in Communication and Critical/Cultural Studies; De Gruyter Handbook of Digital Cultures; Culture, Theory, and Critique; and elsewhere. He is currently Visiting Assistant Professor of Rhetoric and Communication Studies at the University of Richmond.

Email: mac.ogden@gmail.com

Conflicts of interest
None declared

Funding
None declared

Article history
Article submitted: 16/05/25
Date of original decision: 26/09/25
Revised article submitted: 29/09/25
Article accepted: 29/09/25