Communication & Media

Algorithmic Intimacy: How TikTok Teaches Gen Z to Feel

TikTok's algorithm doesn't just show Gen Z content they likeโ€”it shapes how they express emotion, perform vulnerability, and construct intimate connections. Emerging research reveals that 'algorithmic intimacy' is a new form of platform-mediated affect that blurs the line between authentic feeling and performed engagement.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

Something new is happening in how young people express emotion online. On TikTokโ€”the platform that defines Gen Z's media landscapeโ€”emotional vulnerability has become a genre. Users share therapy insights, mental health diagnoses, relationship breakdowns, grief, and existential anxiety in short video format, set to trending audio, formatted for algorithmic amplification. The content is simultaneously deeply personal and expertly performed, raw and rehearsed, intimate and public.

Communication scholars have begun to theorize this phenomenon as "algorithmic intimacy"โ€”a form of emotional expression that is shaped, cultivated, and mediated by platform algorithms in ways that blur the boundary between authentic feeling and performed engagement. The concept challenges traditional models of self-disclosure (which assume a private self choosing to reveal itself in a public context) by suggesting that the algorithm itself participates in the construction of the emotional self.

Defining Algorithmic Intimacy

Sabrina, Danu, and Arindra (2025) examine how algorithmic curation on platforms such as TikTok, Instagram, and X shapes the emotional expressions and perceived intimacy of Generation Z. While Gen Z's digital fluency and identity performance have been widely discussed, less is known about how emotions are cultivated and mediated through algorithmic processes.

The study's theoretical contribution lies in distinguishing algorithmic intimacy from two related but distinct concepts:

Self-disclosure (traditional communication theory): An individual reveals private information to another individual in a dyadic relationship. The motivation is relationship-building; the audience is specific.

Parasocial intimacy (media studies): A viewer develops a one-sided emotional connection with a media personality. The motivation is entertainment or identification; the relationship is asymmetric.

Algorithmic intimacy differs from both because the algorithm is an active participant in the emotional exchange. It selects which emotional content reaches which users, amplifies expressions that generate engagement, and creates feedback loops where the emotional expression is shaped by its algorithmic reception. The user does not simply choose to be vulnerable; the algorithm rewards vulnerability with visibility, creating an incentive structure that cultivates specific forms of emotional performance.

The Aesthetic Dimension

Hussain, Aslam, and Imran (2025) extend the analysis to aesthetic self-presentation. In the age of algorithmically curated social media, beauty standards are increasingly co-produced by platform design and AI-driven filters. Their study investigates the impact of AI-powered beauty filters and recommendation systems on body image and aesthetic preferences among users of TikTok, Instagram, and Snapchat.

The connection between beauty filters and algorithmic intimacy is this: the algorithm does not distinguish between emotional authenticity and aesthetic performance. A TikTok video of someone cryingโ€”whether the tears are genuine, strategically timed, or filter-enhancedโ€”receives algorithmic treatment based on engagement metrics, not authenticity. The platform's inability (or unwillingness) to distinguish genuine emotion from performed emotion creates an environment where the distinction itself becomes irrelevant to the creator's strategic calculation.

For Gen Z users who have grown up within this system, the question "Am I performing emotion or actually feeling it?" may not be meaningful. The algorithm has created a context where emotional expression is always-already performative, because it is always-already addressed to an audience (the algorithm and its human recipients) rather than expressed in a private context.

Identity Construction in Algorithmic Environments

Rustamova (2025) provides the theoretical framework for understanding how algorithmically constituted audiences shape identity formation. The paper investigates the formation and operation of algorithmic audiences within platformized media environments, focusing on how processes of identity, influence, and power intersect to shape audience behavior.

Applied to emotional expression, the algorithmic audience concept reveals that Gen Z users are not expressing emotion to a generic "public" but to an algorithmically curated audience that has been assembled around predicted engagement patterns. The algorithm has already determined which users are likely to engage with emotional contentโ€”and it has assembled an audience that validates, amplifies, and reciprocates emotional expression. The intimacy is real, but its conditions of possibility are algorithmic.

The Historical Trajectory

Tuncay (2025) places algorithmic intimacy within the broader trajectory of AI integration into social media. Through a bibliometric analysis of 996 Scopus-indexed documents from 2014 to 2024, the study maps how AI has reshaped social media research and practice across content personalization, misinformation detection, and automated user interactions.

The study's findings on platform-specific trends are relevant to algorithmic intimacy: TikTok and short-form video platforms are gaining increasing research attention, signaling a shift in how users engage with algorithmically curated content. The growing role of AI in content curation, automated moderation, and chatbot technology has created new dimensions of user interaction that extend to emotional expression. The study also identifies emerging concerns around filter bubbles, algorithmic bias, and deepfakesโ€”all of which intersect with how emotional content is curated and amplified on platforms.

For a generation that has never known a non-algorithmic information environment, these AI-driven transformations in social media are not abstract concerns but the ambient conditions within which emotional expression develops. The boundary between feeling and performing, between authentic and strategic, between human and algorithmic emotional expression may be genuinely different from previous generations' experience.

Claims and Evidence

<
ClaimEvidenceVerdict
Platform algorithms shape emotional expression, not just distribute itSabrina et al. (2025): algorithmic feedback loops cultivate specific emotional performancesโœ… Supported
Beauty filters affect body image and self-presentation among young usersHussain et al. (2025): documented impact on aesthetic preferencesโœ… Supported
Algorithmic audiences are qualitatively different from self-assembled publicsRustamova (2025): audiences are algorithmically constituted with engagement-prediction characteristicsโœ… Supported (theoretical)
Gen Z distinguishes between authentic and performed emotion onlineNo clear evidence; the distinction may be less meaningful in algorithmic environmentsโš ๏ธ Uncertain

Open Questions

  • Is algorithmic intimacy psychologically harmful? If emotional expression is shaped by engagement incentives rather than therapeutic or relational needs, does algorithmic intimacy increase emotional wellbeing (through connection and validation) or decrease it (through performance pressure and inauthenticity)?
  • How does algorithmic intimacy interact with clinical mental health? When TikTok users self-diagnose with ADHD, autism, or BPD based on algorithmically curated content, is this empowering (reducing stigma, enabling access to information) or harmful (promoting pathologization of normal variation)?
  • Can platform design be reformed to support authentic emotional expression? Would chronological feeds, reduced engagement metrics visibility, or content warnings on emotionally manipulative content create healthier emotional environmentsโ€”or would users simply migrate to less-regulated platforms?
  • How do cultural contexts shape algorithmic intimacy? Emotional expression norms vary dramatically across cultures. Does the algorithm homogenize emotional expression toward a global (implicitly Western) norm, or does it adapt to cultural context?
  • Implications

    Algorithmic intimacy represents a genuinely new phenomenon in communication: not merely the digitization of emotional expression but its algorithmic constitution. For communication scholars, this requires extending affect theory, self-disclosure theory, and identity theory to accommodate the algorithmic mediator as an active participant in emotional life.

    For platform designers, it raises questions about the ethical responsibilities of systems that shape how young people experience and express emotion. For parents, educators, and mental health professionals, it demands literacyโ€”not just media literacy but emotional literacy that includes understanding how platforms shape the emotional landscape within which young people develop their sense of self.

    References (4)

    [1] Sabrina, A., Danu, R.T., & Arindra, F. (2025). Algorithmic Intimacy: How Platform Curation Shapes Emotional Expression in Gen Z. KnE Social Sciences, 10(31), 20376.
    [2] Hussain, B., Aslam, S., & Imran, A. (2025). Manufacturing Beauty: How AI and Social Media Are Redefining Aesthetic Norms. Acta Psychologica, 254, 105734.
    [3] Rustamova, N.R. (2025). Algorithmic Audiences: Navigating Identity, Influence, and Power in the Age of Platformized Media. Journal of Communication and Media Studies, 3, 09.
    [4] Tuncay, N. (2025). The 10-Year Shift: How AI Reshaped Social Media from 2014 to 2024. Proc. European Conference on Social Media, 12(1), 3592.

    Explore this topic deeper

    Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

    Click to remove unwanted keywords

    Search 8 keywords โ†’