My FYP is Broken! Tiktok, Algorithmic Dissonance and Scopophilia
Examples of algorithmic dissonance point to the complicated nature of agency, consent and visual pleasure in contemporary social media.
Content note: This article contains discussions of eating disorders.
Algorithmically sorted feeds such as Tiktok´s For You Page (FYP) lend themselves to instances of algorithmic dissonance, or disparities between a user´s self-concept and their “algorithmic representation”. If, like me, you have mourned the loss of your Tiktok FYP to what feels like an incessant stream of conventionally attractive women, attempted to conserve a heavily curated feed by being conscious of all likes and follows, or tried to avoid harmful content with the help of the “not interested” button, you may be familiar with feelings of algorithmic dissonance. While it seems naïve to expect everchanging algorithms and users to coalesce at all times, examples of algorithmic dissonance point to the complicated nature of agency, consent and visual pleasure in contemporary social media.
As is known, Tiktok´s algorithm is said to tailor feeds using “trace data” in the form of “clicks, likes, follows”, shares, pauses and most importantly, watch time. The Tiktok algorithm is premised upon an assumption that equates watch time to an interest in viewing similar content, even if the reasons behind this behavior are often left unexplored.
Interestingly, the app´s focus on measurable “viewing behavior” echoes an earlier turn in broadcasting organisations from citizen to consumer audiences. Taking place in the late 90s, a shift occurred that abandoned ideological considerations of what the audience needed for commercial approaches that prioritized what the audience watched, as reflected in the metrics of “audience flow”, “repeat-viewing” and “channel loyalty”. This change towards audience metrics was based upon a belief that “what viewers say or feel they would like to watch” is often different from “what they watch in practice”.
Although today the Tiktok algorithm is highly attuned to differences in genre and themes, its analogous focus on measurable audience behavior is clear, and here lies one of the central paradoxes of the FYP. Grounded in the assumption that “algorithmic specificity gives rise to the perfect consumption experience”, Tiktok claims to “personalize media-consumption” while “eliminating any opportunities for true personalization”. In this context, algorithmic dissonance seems inevitable as long as users are denied agency over their algorithmic representation.
Similarly, the FYP has been critiqued for its incapacity to successfully censor emotionally triggering content. For example, pro-eating disorder content famously finds its way into the feeds of those who are identified as affectively invested in the former, revealing the problematic applications of metrics such as watch time. When operating in isolation, a focus on user behaviour unfairly tasks the (sometimes underage) viewer with the responsibility of regulating harmful and triggering content.
Yet, while it is imperative that legislating bodies, social media platforms and mental health professionals control access to triggering content, the latter is not easy to delimit. For instance, in the United Kingdom, the Online Safety Bill drafted to grant media platforms “legal duties to minimise harmful content” fails to take algorithms into account or consider the “impact of diet and fitness content” on those suffering from eating disorders. Admittedly, the task of delimiting harmful content becomes almost impossible in a cybernetic space infiltrated by neoliberal ideas of self-discipline that disguise classist, racist and fatphobic concerns with cleanliness and exercise as self-betterment. Masquerading as “wellness” or “lifestyle” content, pro-eating disorder messages are elusive and legitimized by a corrupt societal obsession with health.
While fully exploring the psychological drives that push users to consume triggering content lies beyond my training and the scope of this article, it may be productive to analyze watch time through the lens of scopophilia. If Tiktok´s algorithmic feeds are co-produced and guided by our interactions, then I am at least partly responsible for reducing my FYP to a stream of conventionally attractive women, even when doing so isn´t beneficial nor at first glance interesting to me. It is important to note that there may be a structural predisposition for this particular example of algorithmic dissonance to do with the fact that algorithms are infused with the biases of their creators, which sometimes converge with those of their users. Either way, both can be thought of in relation to Mulvey´s account of scopophilia.
Inaugurating half a century of discussions about the “male gaze”, Laura Mulvey wrote about our fascination with the screen and its contents in her 1975 article “Visual Pleasure and Narrative Cinema”. Here she used the Freudian concepts of scopophilia (the pleasure in looking or “using another person as an object of sexual stimulation through sight”) and identification to describe the unique appeal of film. Some of her conclusions are interesting when extended to Tiktok. For Mulvey, scopophilia consists of “taking other people as objects” through a “curious gaze” in which the viewer/user is given the “illusion of looking in on a private world” while repressing their own exhibitionism. In a similar way, Tiktok videos often claim a casual and authentic nature, supposedly granting us access into a creator´s private world, while allowing us to project ourselves — and perhaps also our desire to be seen — into their online persona. While many Tiktok users prefer taking on a viewing role, not posting any content and thus protecting themselves from the objectifying gaze of others, the very act of watching Tiktoks may to some extent be predicated upon a repressed exhibitionism.
Another notable convergence is the way in which conventionally attractive female Tiktok creators are not only presented as desirable for the viewer, but also for the “characters within the screen story”, or in this case those who make up the private world we claim to have access to (here Tiktok influencer Alix Earle comes to mind). Together with an individual scopophilic drive comes speculation about the effects of subconscious desire on someone´s relationships, friendships and interactions with the world— what must it feel like to be Alix Earle?
Indeed, Mulvey links scopophilia to identification, claiming that since film focuses “attention on the human form”, it encourages a “fascination with likeness and recognition” perhaps symbolized by the correspondence between screen and mirror. For Mulvey, this bodily fascination is gendered and may be shown by the fragmentation of the “female form” (biological determinism aside) in close-ups of women´s legs and face. Not only is the screen/mirror comparison applicable to the Tiktok cyberspace (phone screens become mirrors when the inside camera is used) but so is identification with the female figures of societally informed scopophilia, even if this identification is ultimately treacherous — see more or less ironic comments about Alix Earle being “just like us”. Thirdly, it feels necessary to recall Tiktok´s propensity to exalt different body features as ideal through reoccurring trends. Body parts to have joined the app´s conveyor belt of objectification include noses, side-profiles, lips, stomachs, hips, eyes, eyebrows and cheekbones.
Whether it be the result of a broadcasting culture which privileges observable behaviour and audience metrics, a risk for those exposed to triggering and harmful content which transcends the app, or a case of subconsciously co-produced scopophilia fuelled by identifications with aesthetically exalted others, instances of algorithmic dissonance point to the complicated and interconnected nature of agency, consent and visual pleasure in contemporary social media — it is not just your FYP that is broken!