How algorithmic visibility shapes social, cognitive and teaching presence in educational TikTok comments

This research compares interaction structures in comments posted on TikTok, a platform whose content visibility and dissemination are algorithmically mediated, and derives theoretical and design implications for generative artificial intelligence. Using the Community of Inquiry (CoI) model, the study compares the configuration of social, cognitive, and teaching presence in interactions generated by Ibero-American edutokers with varying levels of visibility. A cross-sectional comparative design is employed, contrasting three different videos (and their associated comment threads) per creator, utilizing deductive coding and relational network analysis. The results show that comments on creators’ first videos are dominated by socially oriented participation (open communication, group cohesion, and emotional expression), whereas cognitive-presence categories are comparatively more prevalent in comments on creators’ most recent videos. Comments on the most-viewed videos form larger and more dispersed networks of associations, in which cognitive categories co-occur less cohesively than in comments on the most recent videos. The most cohesive co-occurrence patterns link exploration, integration, and resolution, rather than reflecting the volume of participation. The findings suggest design considerations for AI-assisted learning environments in which generative AI may amplify interaction while requiring pedagogical mechanisms that preserve conceptual continuity. As a cross-sectional comparison of different videos rather than a repeated-measures panel of the same commenters, these patterns should be read as associative rather than developmental, and generative AI systems were not examined directly in this study.