The moment you open TikTok, your For You feed comes alive. Powered by its recommendation system, it delivers content that almost perfectly matches your interests. That’s the magic of this type of platform: there is no one ‘For You’ page, while different people might come across some of the same standout videos, each person’s feed is unique and tailored to that specific individual (TikTok,2020).
The personalization relies on an accurate user profile, created through algorithmic personalization, which projects to the users their “algorithmic identity – an identity formation that works through mathematical algorithms to infer categories of identity on otherwise anonymous beings” (Cheney-Lippold, 2011, p. 165 as seen in Hearn, 2017, and Bhandari and Bimo, 2022). Further empirical research underscores the reach of such profiling. One example is the study performed by Kosinski, Stillwell, and Graepel in 2013 found that Facebook Likes can be used to automatically and accurately predict a range of highly sensitive personal attributes as sexual orientation, personality traits, intelligence, age, gender, substance abuse, religion, political views, ethnic origin, relationship status, and many others (Kosinski, Stillwell and Graepel, 2013, p. 5802).
The predictions made by these algorithms are perceived positively by platform users, since they can accurately provide personalized content, able to represent the varied, fluid interests and identities of a person (Lee et al., 2022). Indeed, many users have stated that the accuracy of such algorithms are the main reason for either their initial interest in or continued use of TikTok (Bhandari and Bimo, 2022). Conversely, other research has also indicated that when personalization is reduced, users’ engagement drops (Dekker et al., 2025). Hence, the algorithmic personalization that powers the For You page is essential for the users’ experience.
TikTok, unlike other platforms, exploits the algorithm prominence by centering it and by actively incentivizing the user to engage with it. Which means that individuals are not only fully aware of their own relationship with the algorithm, but they are also actively trying to influence its output by adapting their own behavior (Bucher, 2017). People think that the algorithm learns who they are from ordinary browsing and micro-signals – liking, quick swipes, checking comments- and uses them to make predictions about what content they will want to see next. Therefore, they will purposefully engage in algorithmic resistance behavior by interacting with some creators or by withholding clicks from topics they do not want more of (Lee et al., 2022), aiming to shape their feed to reflect their actual or ideal self.
Across multiple studies as those of Lee et at (2022) and Bucher (2017), it was discovered that TikTok’s algorithm not only performs recommendational duties but also works as an emotional conversational counterpart. Users have reported experiences of being understood, laid bare, comforted, and even disturbed by their exposure to their own reflected self in the feed.
Additionally, as Lee et at (2022) note in their paper, TikTok’s algorithmic personalization participates actively in the construction of the user’s self-awareness, perceived social roles, and aspirations. Indeed, they argue that platforms’ recommendation systems dynamically evolve alongside the user, allowing them to “create, remove, and refine new facets to represent the self over time.” This reciprocal process creates a loop where humans both construct and are constructed by algorithmic representation; the algorithm learns from data related to engagement, while humans learn from algorithmic representation to assess their own identity notions.
Similarly, Karizat et al. (2021) observe that TikTok’s For You page facilitates co-produced knowledge of identity as users’ activities determine how “the algorithm knows” the users, but those same algorithmic categorizations reflect “how users come to make meaning about who they are and what kind of social group they belong to.”
Thus, under these conditions, the FYP becomes a collaborative game, a feedback loop, in which the users train the feed, and the feed, in turn, trains users’ tastes and performances.
So…does the For You page truly reflect identity? Partially. The FYP can be seen as a mirror; however, the image that it reflects comes from the user’s own training of the system. This reflection is anything but objective—it is instead informed by the algorithmic logic of engagement, its own biases, and our own activity within the app. What passes for self-recognition is instead a function of both human and computational activity, where certain elements are highlighted, and others obscured or erased. The algorithm learns about the user through the clicks and pauses, but also instructs them to become something else, honoring certain ways of attention and expression over others.
B.BIAS 