Bluesky study: Algorithmic feeds reshape how 4M users write online

Researchers from academia analyzed a massive longitudinal dataset of 235 million posts from 4 million Bluesky users to investigate how sustained exposure to algorithmic feeds influences the language people use. Drawing on Cultivation Theory, the team conducted a quasi-experimental study matching an initial pool of 368,513 users who engaged with one of three curated feeds—News, Science, and Blacksky (a community-focused feed)—against a control pool of 2,001,915 active users who did not engage with any of those feeds. They measured linguistic evolution across three dimensions: lexico-semantics, psycholinguistics, and topics.

The results show that feed exposure leads to significantly greater stylistic accommodation (adopting the feed's writing style), semantic alignment (using similar vocabulary), and register formalization (more structured language) compared to controls. Notably, the Blacksky feed produced the deepest psycholinguistic restructuring, with significant shifts in cognitive processing, affective expression, and pronoun use, while News and Science effects were largely confined to register and topical focus. Regression models revealed that reposting is the most consistent predictor of linguistic convergence across all feeds, whereas posting and bookmarking show feed-dependent effects that differ by more than fourfold. The findings extend Cultivation Theory beyond belief formation to linguistic behavior, highlighting how algorithmic feeds function as persistent linguistic environments that gradually shape what and how users write online.