Algorithmic Bias and Information Diversity: Formation Mechanism and Countermeasure Path of Filter Bubbles in Social Media

Driven by new media technology, the algorithmic recommendation mechanism of short video platforms has enhanced user stickiness through personalised content matching, but also triggered the systematic erosion of information diversity by filter bubbles. Drawing upon relevant academic literature, this study employs a literature review methodology to analyze the symbiotic relationship between algorithmic bias and the information cocoon phenomenon. It further explores the mechanisms through which filter bubbles influence users' cognitive perceptions and the content ecosystem, as well as corresponding countermeasure strategies It is concluded that algorithmic recommendation reinforces user preferences through "data dependence + commercial logic", leading to information narrowing and value illusion; although users can break through part of the closure through active search, they need to rely on the synergy of technology optimisation, media literacy enhancement and policy regulation. Existing research still exhibits gaps in exploring cross-generational user differences, the impact of AIGC technology, and long-term social effects. Future studies ought to construct an algorithmic ethical framework that integrates technical rationality and value rationality.