From friction to alignment: Algorithmic imaginaries and the Trump autocomplete controversy on Reddit

Employing topic modelling, this study analyses a collection of 157 submissions and 10,032 comments from Reddit surrounding users’ perceptions of censorship and bias by Google's autocomplete function following the Trump assassination attempt in July 2024. The discussions are examined through the framing of ‘friction of relevance’, operationalised as the tension between the expectations of individual users and algorithmic moderation by search engines, the strategies users employ to navigate it, and the outcomes when the friction is not resolved. Results reveal users experienced this friction by auditing systems and debating the nature of the moderation. However, the opacity of the search infrastructure, combined with pre-existing algorithmic imaginaries, shaped their views on the technical error, often serving as barriers to resolving this friction. When the friction was not resolved, users saw this as validation of these imaginaries, particularly in the presence of ‘alternative algorithmic imaginaries’, involving perceptions of political bias, censorship and conspiracy theories. This led users to explore and promote alternative search engines in a quest for ‘infrastructural alignment’, believing their individual needs were no longer met. The results show that infrastructural opacity acts as a catalyst for distrust and conspiracy theories, suggesting that explicit communication from platforms is crucial, especially in times of crisis.