Ambivalent algorithms, fragile connectedness: How algorithm awareness facets shape responsiveness and insensitivity on TikTok
2026-08-29
This study examines how distinct facets of algorithm awareness are associated with perceived algorithm responsiveness (PAR), perceived algorithm insensitivity (PAI), and social connectedness on TikTok. I draw on the algorithm responsiveness process and also treat algorithm awareness as four distinct dimensions (content filtering, automated decision‑making, human–algorithm interplay (HAI), ethical concerns) to examine how each dimension is differentially associated with PAR and PAI and, in turn, with feelings of inclusion and belonging. Using survey data from 305 TikTok users and structural equation modeling, I show that user‑agency‑oriented dimensions, particularly filtering and HAI, are indirectly associated with connectedness via higher responsiveness and lower insensitivity. By contrast, automation operates as an ambivalent structure: it boosts responsiveness yet is negatively associated with connectedness, marking a boundary condition for personalization. Platforms should pair automation disclosures with clearer user recourse, traceable feedback, and human review in high-stakes contexts.