CoChat
Sign Up
Cited by 4Ethics and Social Impacts of AIGender, Feminism, and MediaLGBTQ Health, Identity, and Policy

The DOI matches its Crossref record; 13 of the 16 references it cites were found in a registry or an index; none of the 20 sources its claims cite is checked for support yet.

CRUISING TIKTOK: USING ALGORITHMIC FOLK KNOWLEDGE TO EVADE CISHETERONORMATIVE CONTENT MODERATION

Open access
Monea, Alexander·2023·AoIR Selected Papers of Internet Research
DOI 10.5210/spir.v2023i0.13466
Open this paper in CoChat — free accountOpen PDF

Abstract

This paper examines crackdowns on queer content on TikTok and creative responses of content creators to circumvent biased content moderation and cisheteronormative censorship. The first portion of the paper demonstrates TikTok’s recurrent cisheteronormative biases in content moderation decisions and examines select instances of LGBTQ+ content that has been censored on the platform. It also works to situate this within a broader trend of LGBTQ+ censorship across internet platforms. The second portion of the paper examines how LGBTQ+ TikTok users have built up folk knowledges and intuitive understandings of TikTok’s blackboxed algorithms and opaque content moderation policies, situating this discussion within theories of the ‘algorithmic imaginary’. It catalogs the myriad ways that TikTok users work to circumvent LGBTQ+ censorship on the platform (e.g. by tactically obscuring key words in both speech and text and obscuring body parts and scenes). In the final portion of the paper, I draw on the concept of ‘cruising’ and other constitutive silences of LGBTQ+ existence to show how LGBTQ+ users are particularly well suited to producing folk knowledge about blackboxed algorithms. In closing, I examine the affordances and the limitations of LGBTQ+ users’ approach to navigating platform governance – and content moderation practices more specifically – as well as call for more organized and collective action in search of more permanent changes towards LGBTQ+-friendly platforms.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This paper argues that TikTok’s content moderation recurrently censors LGBTQ+ content in cisheteronormative ways, and that LGBTQ+ users respond by developing “algorithmic folk knowledge” to evade moderation. It theorizes these tactics through the algorithmic imaginary and queer “cruising,” while calling for more organized collective action.

Key findings

  • TikTok has censored LGBTQ+ content including depictions of homosexuality, reporting on homosexual groups, gay rights content, and queerness generally, with LGBTQ+ hashtags suppressed in at least eight languages, including Russian and Arabic; Turkey is cited as the most famous example.
  • TikTok’s explanation that restrictions target “pornographic searches” is said to reify the pornographication/hypersexualization of queer existence; the paper notes biased moderation also affects Black creators and the “ugly,” poor, and disabled.
  • LGBTQ+ TikTok users build folk knowledges and intuitive understandings of blackboxed algorithms, using tactics such as intentionally testing trigger words, A/B testing video iterations, misspelling/mispronouncing terms (e.g., “seggs” or “s*ggs” for “sex”), obscuring body parts/background objects/images, leveraging audiences/hashtags/cross-platform links, and sharing knowledge in comment spaces and forums.
  • The paper argues LGBTQ+ users are particularly well suited to producing algorithmic folk knowledge because of queer experience navigating constitutive silences, visibility/invisibility, and cruising; it proposes “algorithmic cruising” as a lens for a more utopic digital future.

Methods

  • Conceptual/theoretical paper examining crackdowns on queer TikTok content and creative user responses to biased moderation and cisheteronormative censorship.
  • Draws on secondary reports and research, and situates analysis within theories of the “algorithmic imaginary,” “algorithmic gossip,” “algorithmic lore,” and “algorithmic folklore.”
  • Catalogs select instances and user tactics rather than presenting an empirical sample, quantitative analysis, or platform data.

Limitations

  • The paper states it examines the affordances and limitations of LGBTQ+ users’ approach to navigating platform governance and calls for more organized and collective action toward permanent changes.
  • Evident from the text: there is no systematic data collection or quantitative analysis; the argument relies on select instances and secondary sources.
  • Evident from the text: TikTok’s algorithms and moderation policies are described as blackboxed and opaque, so users’ folk knowledge is characterized as speculative, anecdotal, and collectively guessed.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.5210/spir.v2023i0.13466

Cites

13 of 13 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 of 20 cited sources checked · 0 supported · 0 flagged

  • Findingnone given

    LGBTQ+ TikTok users utilize comment spaces and forums on other platforms to collectively produce, collect, archive, and disseminate folk knowledge for LGBTQ+ TikTok users.

    “• Utilizing comment spaces and forums on other platforms to collectively produce, collect, archive, and disseminate folk knowledge for LGBTQ+ TikTok users”
    Quote in the paperNot a substantial claim
  • Findingnone given

    LGBTQ+ TikTok users tactically cover or obscure certain parts of the body, background objects, portions of images, etc.

    “• Tactically covering or obscuring certain parts of the body, background objects, portions of images, etc.”
    Quote in the paperNot a substantial claim
  • Findingnone given

    LGBTQ+ TikTok users engage in motivating misspelling and mispronunciation of words understood to trigger censorship, such as 'seggs' or 's*ggs' in place of 'sex'.

    “• Engaging in motivating misspelling and mispronunciation of words understood to trigger censorship (e.g.'seggs' or 's*ggs' in place of 'sex')”
    Quote in the paperNot a substantial claim
  • Findingreasoning

    LGBTQ+ TikTok users continually A/B test the platform's content moderation by posting multiple iterations of the same videos with slight alterations to see which elude the algorithm's content moderation.

    “• Continually A/B testing the platform's content moderation by posting multiple iterations of the same videos with slight alterations to see which elude the algorithm's content moderation”
    Quote in the paperNot a substantial claim
  • Findingnone given

    LGBTQ+ TikTok users leverage specific audiences, hashtags, and cross-platform links to boost LGBTQ+ content.

    “• Leveraging specific audiences, hashtags, and cross platform links to boost LGBTQ+ content”
    Quote in the paperNot a substantial claim
  • It may state two separate claims as one. 6
  • Probably not a substantial claim. 3

  • It may state two separate claims as one.

    Findingreasoning

    LGBTQ+ TikTok users intentionally use language, keywords, hashtags, and images that they anticipate triggering censorship and catalog TikTok's responses.

    “• Intentionally using language, keywords, hashtags, and images that they anticipate triggering censorship and cataloging TikTok's responses”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Premisecitation

    TikTok's biased content moderation extends to many other socially marginalized groups, including Black content creators, as well as the 'ugly,' poor, and disabled.

    “It is worth noting that TikTok's biased content moderation extends to many other socially marginalized groups, including Black content creators (Ghaffary 2021;Rosenblatt 2021), as well as the 'ugly,' poor, and disabled (Biddle, Ribeiro & Dias 2020).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Ghaffary 2021

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Rosenblatt 2021

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Biddle, Ribeiro & Dias 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • It may state two separate claims as one.

    Premisecitation

    The capacity of queer folx to navigate constitutive silences is paralleled by the ability to navigate visibility and invisibility in public spaces, particularly as evidenced by 'cruising' wherein queer folx navigate precarious but coded acts of public disclosure of their sexuality in search of erotic partners.

    “This capacity is paralleled by the ability to navigate visibility and invisibility in public spaces, particularly as evidenced by 'cruising' wherein queer folx navigate precarious but coded acts of public disclosure of their sexuality in search of erotic partners (Bullock 2004;Humphreys 1970;Leap 1999).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Bullock 2004 source
    • cites Humphreys 1970

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Leap 1999

  • Probably not a substantial claim. The paper may give no reason why it holds. It may state two separate claims as one.

    Interpretationreasoning

    TikTok users are concerned with undue censorship of their content and invest time and energy into avoiding having their content censored, deprioritized, demonetized, or otherwise shadow banned.

    “Within this context, TikTok users are rightfully concerned with the undue censorship of their content and invest their time and energy into avoiding having their content censored, deprioritized, demonetized, or otherwise shadow banned.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Premisecitation

    TikTok has censored depictions of homosexuality, reporting on homosexual groups, content promoting gay rights, and content promoting queerness in general in a number of conservative countries, with Turkey as a notable example.

    “It has censored depictions of homosexuality (e.g., holding hands, touching, kissing), reporting on homosexual groups, content promoting gay rights, and content promoting queerness in general in a number of conservative countries -Turkey being the most famous example (Hern 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Hern 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • It may state two separate claims as one.

    Premisecitation

    TikTokers have had to rely on repeated attempts, observed patterns, and gut feelings to figure out how the algorithm works, how to please the platform to facilitate their visibility, and how to have their popularity grow.

    “As Abidin notes,”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Abidin 2020 source

  • Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.

    Interpretationreasoning

    TikTok's argument that many censorship instances were due to restriction of hashtags associated with 'pornographic searches' reifies the pornographication and/or hypersexualization of queer existence that is frequently at the center of cisheteronormative censorship.

    “TikTok has argued that many of these instances were due to their restriction of hashtags associated with 'pornographic searches,' but this simply reifies the pornographication and/or hypersexualization of queer existence so frequently at the center of cisheteronormative censorship.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationreasoning

    Algorithmic folk knowledges require similar capacities to navigate speech and silence, visibility and invisibility, with increasingly high stakes for participating in everyday digitally connected life.

    “Algorithmic folk knowledges require similar capacities to navigate speech and silence, visibility and invisibility, all with increasingly high stakes for participating in everyday digitally connected life.”
    Quote in the paperNot a substantial claim
8 of 8

Citations

Cites

Flagged first, then by citations

3 of the 16 references couldn’t be matched to a known paper.

Research works printed but not linked: 2

  1. Blunt, D., & Stardust, Z. (2021). Automating Whorephobia: Sex, Technology and the Violence of Deplatforming: An Interview with Hacking//Hustling. Porn Studies, 8(4), 350- 366.
  2. Bullock, D. (2004). Lesbian Cruising. Journal of Homosexuality, 47(2): 1-31.

Cited by

Sorted by citations

Related

Metadata

Authors
Monea, A.
Affiliations
George Mason University (US)
Corresponding author
Monea, Alexander
ORCID
Monea, Alexander
Venue
AoIR Selected Papers of Internet Research
Publisher
University of Illinois Libraries
Published
2023-12-31
ISSN
2162-3317 (ISSN-L)
DOI
10.5210/spir.v2023i0.13466
OpenAlex
W4392406478
Semantic Scholar
4a2405056230a301a5e68c919f12a39073a7c600
Cited by
4
References
16
Citations by year
2026: 2 · 2025: 2
FWCI
0.28
Type
journal article
Language
en
Open access
Locations
AoIR Selected Papers of Internet Research (publishedVersion)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Gender Studies › Gender, Feminism, and Media
  • Social Sciences › Psychology › Social Psychology › LGBTQ Health, Identity, and Policy
Keywords
TikTok, content moderation, cisheteronormativity, algorithmic imaginaries, black-box algorithms, circumvention strategies, cruising, platform governance

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review