Year: 2025
Type: article
Study design: Observational Study
Abstract: Many TikTok users report the censorship of ‘sensitive’ content by ‘the algorithm’; this is particularly true of marginalised users. In order to evade perceived censorship, users employ a range of linguistic techniques to obfuscate – hide – their intended meaning. This has received significant media attention, and we complement this by conducting a survey to establish users' motivations for employing these techniques. Our work is informed by linguistic scholarship on self-censorship, anti-languages and platform vernaculars. We conducted a novel survey of 627 UK TikTok users across 2023–2024 (female = 377, male = 224, other = 26) and found that use of obfuscation was relatively low in our sample, and primarily related to the types of content users were posting (historically censored content), rather than acting as a way to establish social identity as we had predicted – though our sample was far from homogeneous on this point. Through a structural equation modelling (SEM) analysis we show that men and people of colour (POC) were significantly more likely to use obfuscation. For POC this is driven partly by positive associations with obfuscation use, suggesting it is seen as a way to be playful with language, as well as evade ‘the algorithm’.
Authors: Eddie L. Ungless, Nina Markl, Björn Ross
Institutions: University of Edinburgh, University of Essex
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Award: G373773752
OpenAlex ID: W4414206587