Theorizing “Stories About Algorithms” as a Mechanism in the Formation and Maintenance of Algorithmic Imaginaries

Abstract

In this article, I report from an ethnographic investigation into young adult users of the popular short-video app TikTok. More specifically, I discuss their experience of TikTok’s algorithmic content feed, or so-called “For You Page.” Like many other personalized online environments today, the For You Page is marked by the tension of being a mechanism of digital surveillance and affective control, yet also a source of entertainment and pleasure. Focusing on people’s sense-making practices, especially in relation to stories about the TikTok algorithm, the article approaches the discursive repertoire that underpins people’s negotiation of this tension. Doing so, I theorize the role and relevance of “stories about algorithms” within the context of algorithmic imaginaries as activating users in sense-making processes about their algorithmic entanglements.

Introduction

Over the course of the last few years, algorithmic content recommender systems have become increasingly common across the web. The popular short-video app TikTok is emblematic for that development. TikTok’s core feature is an algorithmic content feed, the For You Page. Once opening the app, a user is immediately shown this algorithmic feed containing an endless array of short-video clips that TikTok promises will “feel personalized just for you” (TikTok, 2021). The app’s rapid growth in the last couple of years suggests that TikTok is capable of delivering on that promise to at least some degree. In articles with headlines like “How TikTok holds our attention” (Tolentino, 2019), “I spent a week on TikTok and all I got was a new phone addiction” (Spanos, 2019), or “Digital crack cocaine: The science behind TikTok’s success” (Koetsier, 2020), commentators have attempted to explain that success and the appeal of the app. In short, they try to explain TikTok by attending to its algorithm and the affective grip it seems to have on those using the app.

In this article, I confront this interest in the TikTok algorithm, its powers, and the question of how people deal with it on a day-to-day basis. To do so, I report from a larger ethnographic project on the TikTok use of young adults based in the United Kingdom. Drawing on the “algorithms as culture” approach (Seaver, 2017), I understand the TikTok algorithm not as a bound technological object but rather as a cultural artifact enacted in everyday processes of meaning making. Attending to media technologies by examining how they are figured and domesticated as cultural artifacts is by no means a novelty (consider for instance the work of Morley, 1986; Silverstone, 1994; or Williams, 1974). However, only more recently has a line of research started to form that investigates algorithms in the nexus of their figuration and encounters with them in daily life (see, for example, Bishop, 2019; Bucher, 2017; or Siles et al., 2020).

Contributing to this tradition of media studies and scholarship on algorithmic imaginaries specifically, this article highlights the role that stories play in such processes of imagining and interacting with algorithms in everyday life. “Stories about algorithms” will be theorized in broad terms as the ways in which ordinary users share among each other their lived experience of algorithms. Such stories, I argue, don’t make people more knowledgeable about algorithms but are crucial mechanisms which reactivate people in those sense-making processes about invisible algorithms upon which algorithmic imaginaries are formed and maintained.

The article opens with a review of current debates on algorithmic imaginaries, arguing that there is a gap in scholarship in terms of understanding how algorithmic imaginaries of ordinary users are organized on a social level. To fill this gap, I introduce the idea of “stories about algorithms” as the way in which ordinary users share their lived experience of algorithms among each other. Following a brief discussion of the ethnographic project this article reports from, I illustrate this aspect of “stories about algorithms” through the case of TikTok. I outline the overall imaginary of the TikTok algorithm that I encountered during my fieldwork and discuss two cases of stories that I observed shaping it. First, I present the case of “if you see this . . .” videos as stories that reactivate users in their sense-making about the TikTok algorithm as a mechanism of digital surveillance. Second, I outline the case of “TikTok rest area” videos as stories through which people create awareness for the TikTok algorithm and its “addictive” nature. Doing so, the article will be closed with a reflection on the role of stories as a relevant form of knowledge and support mechanism in the context of people’s algorithmic imaginaries and entanglements.

Scholarship on Algorithmic Imaginaries

Conceptualizing “Algorithmic Imaginaries”

Studying algorithms as culture, as “unstable objects, culturally enacted by the practices people use to engage with them” (Seaver, 2017, p. 5), has become a preferred approach by many social scientists in recent years. It is an approach that opposes understanding algorithms solely in technical terms. Such technical terms define algorithms as mere sets of encoded procedures and instructions producing defined outputs, like that of a content recommendation. Instead, social scientists look at algorithms in their social embeddedness (see, for example, Bucher, 2018; Kitchin, 2016; or Gillespie, 2014). The “algorithms as culture” approach defines them not through their inside, so to speak, but rather along their outside edges, the way in which they are integrated into society and the lives of individuals encountering them.

In this article, I follow the “algorithms as culture” approach and especially draw on the idea of the “algorithmic imaginary” developed in it. At its core, scholarship on algorithmic imaginaries is interested in understanding what people—be they professional content creators, ordinary users, or larger publics—know and think about algorithmic systems. However, where this scholarship shares a common interest in knowledge about algorithmic systems, approaches differ in their concrete theorizations thereof.

Taina Bucher’s (2017) work on the “algorithmic imaginary” is among the most prominent theorizations in the field. Bucher defines the algorithmic imaginary as the “way of thinking about what algorithms are, what they should be, how they function and what these imaginations in turn make possible” (Bucher, 2017, p. 40). Bucher crucially stresses the importance of imaginaries being more than just mental representations of algorithms. Rather, they are to be understood (and studied) as something that is productive. While users might not be able to see algorithms or understand them on a technical level, they nonetheless sense and feel their consequences. It is upon this felt presence of an algorithm that an algorithmic imaginary develops. This imaginary shapes the modality of future interactions and experiences users have with a given algorithm (Bucher, 2017, p. 42).

Scholarship drawing on the idea of “folk theories” differs from approaches linked to the idea of the “algorithmic imaginary.” This difference emerges in that studies of “folk theories of algorithms” are mostly concerned with mapping the variety of different theories that are present within a group of users or larger publics about specific algorithmic systems (compare DeVito et al., 2017; Eslami et al., 2016; Ytre-Arne & Moe, 2021). Doing so, as Ignacio Siles et al. (2020) argue, they often fall short in “showing how theories and imaginaries of algorithms relate to specific sets of action strategies that shape modalities of power and resistance” through which algorithmic entanglements are enacted (Siles et al., 2020, p. 13). Making this observation, Siles et al. fruitfully set out to combine the two concepts, “algorithmic imaginary” and “folk theories,” respectively. Drawing on the former, they stress the constitutive role of people’s imaginations in enacting the specific modality of a given algorithm-user relationship. However, drawing on the latter, they underscore that such imaginations need to be studied as plural. For each algorithm, there exists a discursive repertoire of different theories outlining different possibilities of enacting it in cultural processes of meaning making.

In that context, the work of Sophie Bishop (2019) is vital because it highlights the social organization of such imaginaries and theories. Discussing ethnographic observations from fieldwork with beauty vloggers on YouTube, Bishop introduces the concept of “algorithmic gossip.” She defines this kind of gossip as “communally and socially informed knowledge about algorithms and algorithmic visibility” (Bishop, 2019, p. 2590). Bishop describes a setting in which professional content creators are unable to investigate, on a technical level, how the YouTube algorithm works and promotes their content. In that setting, Bishop observed a discourse of professional creators sharing their individual experiences of the YouTube algorithm among each other to create not a technically but socially valid account of how it can be interacted with (Bishop, 2019, pp. 2602–2603).

In sum, Bucher (2017) has initially conceptualized the “algorithmic imaginary” as emergent from people’s lived and embodied experiences of algorithms. Studies like those of Siles et al. (2020) or Bishop (2019) advance this conceptualization by locating this everyday knowledge about algorithms beyond the domain of personal experience. They highlight how algorithmic imaginations constitute a larger discursive repertoire that is socially organized and maintained. Joining this line of scholarship on algorithmic imaginaries, in the next section I introduce the idea of “stories about algorithms” as a mechanism and way of studying this dynamic of algorithmic imaginaries being constituted both on a social and personal level.

Conceptualizing “Stories About Algorithms”

I define “stories about algorithms” in broad terms as the ways in which ordinary users express and share among each other their lived experience of algorithms. Following Walter Benjamin (1977), I contrast the domain of stories from that of information. Information, as Benjamin argues, only has value in the moment it can verify its claim about reality. Stories, in contrast, speak from lived experience, be it one’s own or shared, experience passed from mouth to mouth (cf. Benjamin, 1977, p. 386). It is this ability of sharing lived experiences, and the guidance and orientation it provides people in their lives, that Benjamin feared lost in societies increasingly organized around the logics of information and news (cf. Benjamin, 1977, p. 390). This development, the loss of stories and their guidance, resonates with the contemporary condition of algorithms and how they come to shape online environments.

Recent studies, like those of Minna Ruckenstein and Granroth (2020) on advertising algorithms, showcase how users often feel alone and ill-equipped in their sense-making when they encounter algorithmic recommendations that are confusing and disturbing to them. While people might know what algorithms are, that they select the content they see online following the logic of personalisation, that kind of information in and by itself does not help them in moments of disturbance. This information doesn’t offer them orientation and support in the pragmatic sense of dealing with algorithms on a day-to-day basis.

Experimental studies, such as those of Motahhare Eslami et al. (2018), show that users who have more information available to make sense of the recommendations they are served by algorithms feel more comfortable about their encounter with them than those who don’t. Such studies underscore that there exist better ways and more information that companies could provide users to help them in their everyday encounters with algorithms. Nonetheless, verifiable and useful information about algorithms remains scarce and impractical to obtain in everyday settings. This speaks for the need to investigate the role that other forms of knowledge, like those of stories, as opposed to domains of information alone, can and do play in algorithm-user relationships.

Analytically, the idea of “stories about algorithms” thus opens a broader horizon in which algorithmic imaginaries can be located and studied as relevant form of knowledge. Thus far, scholarship on algorithmic imaginaries has mostly been concerned with investigating them through methods such as surveys, interviews, or focus group discussions (with notable exceptions in the ethnographic projects of scholars like Bishop, 2019). Scholars have mostly investigated everyday knowledge about algorithms in a research environment. A common way to design such investigations has been to listen to “narrated emotional reactions” from past encounters with algorithms (see discussions in Bucher, 2017; Kennedy & Hill, 2018; Ruckenstein & Granroth, 2020). The work of Bucher (2017) in particular has proven that this, overall, is a very fruitful approach. Listening to people’s narrated emotional reactions, Bucher was able to uncover not only the ways in which people imagined algorithms of different social media platforms but also how they enacted their affective relation to them. Based on their imaginary of an algorithm, like that of YouTube or Facebook, Bucher observed people trying to “train” algorithms by consciously liking certain contents, for example (Bucher, 2017, p. 41).

The work of Bishop (2019), as mentioned, differs from these approaches in having investigated algorithmic imaginaries within the field, in her case as a form of “gossip” unfolding among professional content creators. However, in the context of “ordinary users,” this kind of “gossip” remains understudied and undertheorized. As ordinary users I understand people that primarily sit on the receiver end of recommender algorithms. Such ordinary users, I argue, are more audience-centric than producer-centric. Subsequently, they imagine algorithms from a fundamentally different angle compared to professional content creators for whom algorithms primarily are means of self-promotion, and not a source of pleasure or entertainment. Stressing the importance of Bishop’s work (2019; see also 2020), I thus position “stories about algorithms” in a similar way to her conception of “algorithmic gossip,” namely as informal and bottom-up discourses about algorithms that take shape when ordinary users share their lived experience among each other.

In that sense, this article outlines the need for methodological duality when investigating algorithmic imaginaries. There needs to be an approach of listening to people’s narrated emotional reactions in interviews. However, there also needs to be a more ethnographic investigation of how such imaginaries of algorithms are organized and maintained through stories that extend beyond the horizon of personal experience. Therefore, I draw on Benjamin’s (1977) reflections on “stories” primarily for their heuristic value. I define stories about algorithms in broad terms because, as I will show, the specific forms in which they materialize vary and are only of secondary importance. What is more crucial, on an analytical level, is sensibility for stories, for the ways in which ordinary users share not with researchers but among themselves their lived experience of algorithms. Benjamin’s (1977) reflections on “stories” enable this kind of sensibility. They allow to design a process of uncovering algorithmic imaginaries as a form of everyday and social knowledge in a fuller and more complete way.

In the remainder of this article, drawing on the case of TikTok, I showcase the benefits of this dual approach. Stories, I will argue, do not make people more knowledgeable about algorithms and how they work. However, this does not mean that they are of lesser value as a form of knowledge. Instead, I will show that they play a crucial role in shaping user experiences by creating external impulses and sites of awareness for that which otherwise remains latent: the algorithm. This is why the meaning of “stories about algorithms” can only be understood when they are read in the context of a broader algorithmic imaginary, hence require a dual approach. Following a brief discussion of the ethnographic project the article reports from, I therefore first outline the broader algorithmic imaginary of TikTok that I encountered during fieldwork and then locate the element of stories about the TikTok algorithm within it.

Fieldwork and Methodology

This article uses material from a larger ethnographic investigation into the short-video platform TikTok. Fieldwork for this project started in early 2020 with a 6-month-long digital ethnography of the TikTok platform. During this period, I immersed myself in the rhythms and flows of content that characterized the app (see Schellewald, 2021). Starting in the summer of 2020, I conducted a series of interviews with 30 young adult users of the app, all based in the United Kingdom. Participants were recruited using a promoted tweet that was targeted at Twitter users aged 18–24, interested in TikTok, and based in the Greater London area (the final group of participants largely matched these categories, with only a few people being slightly older and living outside the Greater London area). In public debates, TikTok is often perceived as a “kids app.” However, young adults form the largest group of users on the app in the United Kingdom (compare Loose et al., 2020), which is why my project focused on them. Fieldwork with this group of 30 young adults ended in the summer of 2021.

Overall, the study was concerned with understanding why and how young adults use TikTok in their everyday life, as well as how they interact with the TikTok algorithm. Other than presenting a complete discussion of that ethnography, this article draws on material from it to illustrate the facet of stories as part of the algorithmic imaginary surrounding TikTok. Reporting from this broader project, it is important to stress that in relation to the TikTok algorithm I initially followed the approach of listening to people’s “narrated emotional reactions” (similar to Bucher, 2017, and others). However, doing so, as well as during my fieldwork on the TikTok platform, I frequently encountered TikTok videos that thematized the TikTok algorithm and people’s experience of it. It is based on this observation that I turned to the aspect of “stories” as a social dimension of people’s algorithmic imaginary.

In short, this article discusses the aspect of “stories about algorithms” not holistically but from the particular angle of my broader ethnographic project on the TikTok use of young adults based in the United Kingdom. My data and methodology is thus limited to this specific cultural context, as well as to answering questions about what “stories about algorithms” are and how they interact with people’s algorithmic imaginaries and user experiences of the TikTok app.

The Algorithmic Imaginary of TikTok

TikTok and the Pleasures of Scrolling

To understand how the TikTok algorithm is imagined by “ordinary users” of the app, we first need to understand how the app is used by them. There are many things that one could use TikTok for, such as creating and sharing content, connecting with like-minded people, participating in public discourse, or simply staying in touch with friends and family (many of these uses have been discussed in existing TikTok scholarship by Abidin, 2020; Boffone, 2021; Kaye et al., 2021; Literat & Kligler-Vilenchick, 2021; or Zeng et al., 2021; among others). Yet, none of my participants really used TikTok for these reasons or in such ways. Instead, they mostly just scrolled through the For You Page. Hence, their experience of TikTok was always also an experience of the TikTok algorithm and its workings. For them, being on TikTok was a form of “me time.” They already had media in place which they used for communicative needs like following news and current affairs, participating in public debates, or to stay in touch with friends and family. Apps like Instagram or Twitter were central to people’s communicative routines in these regards. While they appreciated and valued them very much, especially during times of lockdown in which their social life had to be carried out at a distance, there was something that these apps were unable to afford.

The moments in which my participants ended up on TikTok were always moments in which they felt overwhelmed, stressed, anxious, tired, bored, or simply needed cheering up and a break from life in the here and now. Their use of the app wasn’t really planned but more spontaneous. One of my participants, Sunder, described this by explaining that “it’s like you’re just opening the various social media apps that you have and almost are waiting for the right thing to appear to entertain you.” In this article, TikTok needs to be understood in the context of such quests for distraction in everyday life. As another participant, Bea, explained to me:

When I go on TikTok I can just watch videos of cute dogs and stuff. It’s more of an escape than Twitter or Instagram for me, especially right now. It’s just a nice way to avoid life. On TikTok time is kind of paused. It’s like a true 30-minute escape that you can’t leave once you’re there. (Bea)

The reason why TikTok is able to do so effectively is, as Bea told me, because “I think the algorithm knows me pretty well” (Bea). Many participants shared and expressed similar sentiments about the TikTok algorithm, how it was very good at being able to show just the right thing to catch and hold their attention. In that sense, Bea’s comments point to a tension that I noticed in the experience of all my participants. People enjoyed TikTok because it was an app that you could “just” scroll through. However, they were also very aware of being overserved and profiled by TikTok for that purpose. The same was the case for the app’s affective grip, or it’s “addictiveness,” as my participants described it. They at once desired to get carried away by the app, yet were also aware of TikTok trying to keep them engaged for as long as possible.

Becoming Aware of the TikTok Algorithm

The way in which my participants made sense of TikTok as an algorithmic environment revolved around this tension of it being at once a source of pleasure and relaxation, yet it also being a mechanism of surveillance and affective control. My participants integrated TikTok into their everyday routines as a dedicated “feel good space” and a form of “me time.” They had clear expectations, like Adna, who argued that “if I’m on TikTok I don’t want to see anything annoying or unfunny” (Adna). My participants wanted TikTok to be composed of videos that, overall, create a homogeneous textual structure that feels “for you.” However, while people desired such a seamless feed showing only the right type of content, the reality was more complex and messy.

People frequently became aware of and active in relation to the TikTok algorithm and how it curated content for them. Other than personalisation being a solely technological consequence, during my fieldwork I rather encountered it as the product of a continuous process of algorithm-user interactions. In this process, users imagined TikTok as “for you” and resultingly learned how to attune themselves to the TikTok algorithm in a way that enacts a pleasurable experience of “me time.” Subsequently, people’s imaginary of TikTok as an algorithmic environment was shaped in relation to assessing the accuracy of the TikTok algorithm and the content it recommends to be “personalised.” People noticed the algorithm when it showed content they had purposefully tagged with a “like” to see more of it. However, they also mentioned that “sometimes [the algorithm] is a bit trigger-happy” (Judith), overinterpreting behaviors such as liking videos (cf. similar findings in Siles & Melendez-Moran, 2021).

In response to this “trigger-happiness” my participants, over time, started to adjust their behavior on the app. They became much more mindful of which videos they would tag with “like,” how fast they would scroll past something they didn’t immediately enjoy, or who they would follow, for example. These practices weren’t meant to control the For You Page but rather were aimed at nudging the algorithm to stay in line (cf. Bucher, 2017, who observed similar practices for other platforms). It is here that we can see people’s imaginary of the TikTok algorithm at work. The TikTok algorithm was imagined to deliver recommendations “for you” and, subsequently, people enacted their relation and behavior to it accordingly, actively helping to construct the “for you” through their own actions and practices.

While this imaginary was primarily shaped by people’s personal experiences of using TikTok, “external” factors also played a crucial role. During fieldwork, TikTok had been frequent object of public scrutiny in the media (cf. Stokel-Walker, 2021, for an extensive overview of these public debates). Although people were aware about these debates and concerns raised in them, conversations with friends or stories about the TikTok algorithm shared online played a much more important role in shaping their sense-making on the app. Some, like Iris, who had been using TikTok since early 2019, even mentioned to me that they only became fully aware of the TikTok algorithm after seeing videos on their For You Page of people “making jokes about it being this algorithm” (Iris). While not the first realization for others, almost all participants mentioned that they would frequently see such videos on their For You Page. In the rest of the article, I explore discursive resources of this kind—TikTok videos telling stories about the TikTok algorithm—and how they aided people in their affective relations and imagination of TikTok as an algorithmic environment.

Stories About the TikTok Algorithm

Case 1: “If You See This . . .” Videos as Stories About Digital Surveillance

What my participants enjoyed about TikTok, in contrast to other apps, was how easy it was to connect with “ordinary strangers,” people just like them, on the platform. Some even described the feeling of scrolling through the For You Page as “seeing what your friends are up to if your friends were all kind of cool and you didn’t have any obligation to them” (Judith). The For You Page did, in short, appeal to people for how specific it was to their personality, lifestyle, and current life situation. In relation to this pleasurable specificity, there had emerged a specific trend on TikTok. One of my participants, Velta, a recent university graduate that used TikTok a lot while living alone during lockdown, described it as follows:

There is this trend where people are like, “the algorithm apparently shows you creators that have the same interests and same personality traits as you.” And then these videos basically say, “because of this you are also single, a university student, living in London, . . .” and it was all these traits, all of them applied to me . . . and no hashtags or anything. So, I was just really curious how they knew all that. (Velta)

“If your see this . . .” videos open sites of awareness for the TikTok algorithm from within the For You Page. They have a clear meta communicative form (cf. Schellewald, 2021, for a broader discussion of TikTok’s different “communicative forms”). They are communication about communication, or, in other words, stories about the TikTok algorithm. In response to the lived experience of personalisation on the platform, users create videos like Velta described them, videos that attempt to be a sort of living proof that the TikTok algorithm is capable of figuring you out.

Almost all participants mentioned having seen such videos. In the case of Velta, seeing this concrete materialization of the TikTok algorithm’s specificity made her curious and start thinking about how it works. Similarly, Benjamin, a psychology student, mentioned that seeing this type of content often reminds him of debates on digital surveillance:

It was the “if you see this, you’re probably gay, under the age of 25, . . .,” like this, this, this, and it was literally almost like quite uncanny how accurate it got. It made me think about how everyone jokes about like talking to someone about a certain product and then the next ten minutes you are on your phone and it comes up as an ad and you’re like, “omg all the technology is listening.” It reminded me of that, but I wouldn’t say that I felt uncomfortable. I have enough awareness to know that if I’m going onto an app where it has a “For You Page,” I’m going to expect that it tracks certain things. (Benjamin)

Other than mediating new knowledge about the TikTok algorithm, “if you see this . . .” videos rather reactivate thought processes through which one’s algorithmic imaginary develops and had developed previously. In the case of Benjamin, this meant thinking again about the topic of digital surveillance and how he had negotiated it for himself. Yet, of course, not all people felt as comfortable with it as Benjamin did. Some, like Jade, a politics student, talked about seeing one of these videos in the following way:

I actually got scared for a moment and was like “oh wow . . . ok . . . this is way too close” and “maybe China is actually monitoring us” [laughs]. Yeah, I remember that one was just like “what’s going on?” (Jade)

For Jade, such videos were productive of a view that many people held, namely that the TikTok algorithm has both an appealing but also scary accuracy (cf. also the findings of Simpson & Seeman, 2020). While enjoying the For You Page because it is personalized and specific, sometimes the TikTok algorithm also appeared to go too far. It places a video on people’s For You Page that is experienced as being so specific by the person scrolling that it crossed a boundary between personalisation perceived as appealing and personalisation perceived as “too close to home.”

On one side, trends like the “if you see this . . .” video appear thus a response to people grappling with the issue of obtaining verifiable information about the TikTok algorithm. As a video that says, “if you see this . . . then you are,” they provide an opportunity for users to easily verify if the TikTok algorithm is, in fact, accurate. However, I locate them here in the domain of stories because they primarily speak from lived experiences. They are, quite literally, stories told about what people noticed about the TikTok algorithm and now share with other ordinary users. However, this is also because the verifiable information they offer, in and by itself, is not what I found makes them meaningful as discursive objects. As Velta’s example shows, in some sense a “if you see this . . .” video does not really reveal much. Rather, it made Velta wonder more about how exactly the TikTok algorithm is capable of doing what it does.

Instead, stories like those of “if you see this . . .” are meaningful discursive objects by brining to the foreground what otherwise remains latent and at most felt while scrolling, the TikTok algorithm observing and targeting “you.” As Walter Benjamin argued, a story unfolds its full potential only when listened to almost bored by it (Benjamin, 1977, p. 392). Trends like the “if you see this . . .” video underscore that, I found. Their critical potential is less the information they verify than how they contribute to creating from within the For You Page repeated moments of confrontation with the TikTok algorithm. Other than determining how exactly people feel and think about this algorithm, they are external impulses that reactivate exactly those thought processes that are constitutive to the development of algorithmic imaginaries on the personal level.

Case 2: “TikTok Rest Areas” as Stories About Affective Control and Addictiveness

Where the “if you see this . . .” story creates awareness for the TikTok algorithm as a mechanism of digital surveillance, the story told in so-called “TikTok checkpoints” or “TikTok rest area” videos does so for the TikTok algorithm as a means of affective control. As a trend, the “TikTok checkpoint” is less prominent in comparison to “if you see this . . .” videos. In fact, none of my participants had seen content of this kind and I only discovered it during my own fieldwork on the platform. However, although a bit more niche, I argue that as a case these kind of videos nicely showcase and allow to explore the impact that stories about algorithms can have on how users manage their entanglement with captivating online environments like TikTok. More specifically, it nicely highlights how seemingly random and trivial “stories about algorithms” enact a sort of collective support environment that helps people negotiate their entanglement with invisible algorithms.

A “TikTok checkpoint” video, at its core, is designed in a way that opposes the rhythms and styles that are perceived as predominant on the platform. Like the “if you see this . . .” video they, too, stand out and break with the flow of the For You Page. Such videos, as shown in Figure 1 below, normally feature a text annotation stating that the video is a “TikTok rest area” or that one has reached a “TikTok checkpoint.” Some follow up on that introduction with messages stating that everyone is invited to stay for as long as they want. Some videos, furthermore, add more specific comments, like asking people to take a break from scrolling, to put away their phone, or remind people of things like staying hydrated.

Figure 1.

Screenshots of “TikTok rest area” videos form users @tofu_corgi (2020), @riiiiiicola (2020), and @fishie.fish (2019).

In their aesthetic style, “TikTok rest areas” mostly refrain from using any visual filters or effects and compose their video as having a generally calm tonality. All of those which I saw during my fieldwork on the platform were single, unedited shots, lasting for up to 30 seconds or even a minute. The same is true for their audio, which mostly consisted of atmospheric music or ambient sound effects like rainfall noises. In this aesthetic style, “TikTok checkpoints” directly contradict the more flashy, fast paced, and energetic tone that characterizes much of TikTok’s more mainstream content.

In this form, they tell a story, are created in response to the lived experience of the predominant rhythms and speeds that mark the For You Page. Where the dominant logic of the TikTok For You Page is that of an endless and continuous flow, “TikTok rest areas” create speed bumps within that. They invite people to stay, rest, and slow down, all without having to leave the app. Similar in the way that “if you see this . . .” videos don’t tell people how to think or feel about the TikTok algorithm, such “rest area” videos do not determine how long one’s break will be. They simply create a noticeable break from the pace and pattern of the For You Page. They interrupt the flow and require the user to become active, to think about and manage their entanglement with the TikTok algorithm.

In terms of their effectiveness, there are issues at hand. These kind of videos still need to be placed on people’s content feed by the TikTok algorithm to provide an opportunity to break from the app’s captivating flows. Furthermore, there are similar measures that people have often put in place for themselves. For instance, some of my participants, like Tanja, mentioned how they have a time limit on how long they can use TikTok on their phone each day, creating a functionally similar kind of external impulse to regulate one’s affective relation to the TikTok algorithm.

Yet still, as a case, I argue that “TikTok rest areas” highlight the potentialities that sit on the user-end in terms of aiding other people in managing their relation to algorithms like TikTok’s. Their challenge of having to appear on people’s For You Page also being a core aspect of their potential. These videos can potentially turn up on people’s content feeds and open sites of awareness within the very flows of content they combat. Especially in the case of TikTok, this video-based format appears much more experientially engaging than that of similar forms, like that of user comments.

It is not uncommon for people, in the comments of a video on TikTok, to address the current situation through which they ended up seeing that specific video. Lisa, for example, told me that sometimes she goes through the comments of TikTok videos, including when she ends up scrolling through the app late at night. In relation to reading such comments, Lisa explained,

If it connects like to the situation I’m watching the video in . . . so, like, if the comment is “me watching this at 2 am” and I’m up around that time as well, then I’m like . . . it almost makes me feel a bit better [laughs] about being in that state, because it’s like obviously the experience that everyone is having. It’s kind of nice knowing that it has a common effect on people. (Lisa)

Will had told me a similar story. He had stopped using TikTok roughly half a year following our initial conversation, but then picked up YouTube’s short-video service called “Shorts.” Watching videos on YouTube Shorts, he mentioned sometimes seeing similar comments like those Lisa mentioned, such as people stating things like “it’s 3 am, why am I here?.” Asking Will what happens after he reads such comments, he responded by saying:

That would be the end point. That would be the end of a video for me. As in reading that I was like “oh wow, ok . . . this has been 45 mins, I should do something with my life.” (Will)

What is interesting about these interactions between ordinary users through stories is their ephemeral and serendipitous nature. It is within these small moments, the stories through which ordinary users share their lived experience of algorithms, that they enact a support environment. This environment does not take the form of a tight-knit community in which mutuals provide continuous support for each other. Instead, it is a more abstract formation that, so to speak, brings light into darkness. Scattered throughout online spaces like TikTok these small stories enable people to see the TikTok algorithm and themselves in relation to it. In this form, they highlight how without much effort and centralized organization there already exists the potential to have a meaningful impact in algorithmic power dynamics. As the case of Will demonstrates, it appears that sometimes almost all it takes to break free from the affective grip of a recommender algorithm can be a comment posted under a video that poses the question, “why am I here?.”

In sum, awareness for one’s condition of consumption sparked by reading comments or having one’s flow of content interrupted by a “TikTok checkpoint” creates vital impulses. Such stories about an algorithm’s affective grip are speed bumps on the For You Page. If one encounters them, they require a reaction. They require the person scrolling to become an active agent in the algorithm-user relationship. Even if it that activity would just be to skip past the speed bump and keep the flow running. As a case, phenomena like the “TikTok reast area” and comments like Lisa and Will described them do nicely showcase the crucial role that stories about algorithms can play. They create awareness for invisible algorithms and reactivate people in their affective relation to them.

Conclusion

In this article, I have introduced the idea of “stories about algorithms” alongside the case of TikTok and young adult users of the app based in the United Kingdom. As discursive objects, stories about algorithms do not offer their listeners new information about what algorithms are, how they work, or how they can be interacted with. Instead, what they do is make tangible what otherwise remains latent and invisible: the algorithm. Doing so, stories about algorithms (re)activate those thought and reflection processes upon which an algorithmic imaginary takes shape on the personal level. Algorithmic imaginaries being, to reiterate, the structures of everyday knowledge about algorithms that shape how users approach, interact with, and experience algorithms like that of TikTok.

Introducing the idea of stories in this way, I do not argue that they are the sole way in which people’s imaginations of and interactions with the TikTok algorithm were shaped. Quite the opposite, people experienced TikTok as a very unique algorithmic environment, one that, because of its dynamic nature, itself was very noticeable when encountering it. Within such encounters, people’s imaginary of the TikTok algorithm developed most centrally. However, what I discovered during my fieldwork was how stories about algorithms, in their sum, materialized a kind of support environment in which ordinary users created among themselves continued awareness for their shared condition of living with surveillant and captivating algorithms.

Yet, as scholars like Ien Ang (1996) have prominently reminded us in debates on the “active audience,” “it would be utterly out of perspective to cheerfully equate ‘active’ with ‘powerful’, in the sense of ‘taking control’ at an enduring structural or institutional level” (Ang, 1996, p. 117). The same is true and crucial, I conclude, for the concept of the “algorithmic imaginary” and the notion of “stories about algorithms” as I have introduced it. The moments in which my participants were active in relation to the TikTok algorithm were primarily navigational in their form. They weren’t aimed at controlling the For You Page but managing one’s passage through it. These forms of activity mattered to people, not in terms of “powerfulness,” but in allowing them to not feel completely helpless in the face of algorithmic power. It is crucial to remember that, after all, most users engaging with algorithmic environments like TikTok are just ordinary people trying to go about their daily life. The idea of “stories about algorithms,” as I have presented it in this article, offers an analytical entry point to explore and understand this ordinary dimension of algorithms, the imaginaries that surround them, and their organization on a social level.

As mentioned earlier, I only stumbled across these stories during a broader ethnographic investigation of TikTok. In that regard, I have here only provided an initial perspective on the phenomenon, namely in defining “stories about algorithms” and showing how they impact sense-making processes and user experiences of algorithmic environments like TikTok. From the point of view of my participants, who engaged with TikTok in their search for distraction and entertainment, it only mattered that these stories exist, that they helped them navigate the TikTok algorithm and its tensions. Who created and shared these stories, and for what reasons, wasn’t really a concern for my participants. To explore such stories and the support environments they enact from more angles appears thus both very fruitful and important. It does so precisely because it will crucially broaden our understanding of the spaces where algorithms and users meet, how power is negotiated in them, and how these processes of negotiation can be supported effectively.