Abstract
This paper explores the mechanisms of algorithmic linguistic manipulation in virtual discourse across English-, Russian-, and Kazakh-language digital platforms. It treats algorithms not merely as passive tools but as active discourse subjects that curate, suppress, and sometimes generate language in ways that influence public opinion. Drawing on empirical data from Twitter/X, Facebook, Telegram, TikTok, and local Kazakhstani platforms, the study uncovers non-obvious linguistic strategies employed by bots, political actors, and users—including bilingual code-switching, dialectal mimicry, semantic re-packaging, and engagement hacking. Special attention is given to Kazakhstan’s bilingual online environment and evolving digital sovereignty framework, showing how Kazakh and Russian linguistic spaces are algorithmically shaped by both domestic and foreign influence operations. The article integrates theories of language ideology, media control, and algorithmic governance to explain how discourse is engineered in digital spaces. The findings suggest that multilingualism can both complicate algorithmic moderation and serve as a defense against manipulation, depending on how communities adapt. By comparing three linguistic spheres, the article highlights structural asymmetries in content moderation and proposes context-sensitive strategies for preserving authentic digital discourse in the algorithmic age.