Linguaggio tossico e costruzione del consenso su Instagram e Facebook. Un’analisi dei contenuti politici e giornalistici in campagna elettorale

Toxic and uncivil language on digital platforms is a contentious phenomenon, often seen as a potential driver of user engagement and, as such, strategically leveraged to enhance visibility. This study examines the role of toxicity in the content shared by political actors and news outlets on Facebook and Instagram during five Italian electoral campaigns (n = 117,805). Using the open-source Detoxify model to detect textual toxicity, the research addresses three questions: the presence of toxic language, its evolution over time, and its relationship with user engagement. The findings show that, although toxic content represents a small proportion of the total, it tends to generate higher levels of interaction, particularly in political posts. Moreover, toxicity appears to be a contingent rather than structural feature of online political discourse, shaped by specific moments and contexts. These results raise important concerns about the quality of democratic debate online, as political visibility and authority may be strategically reinforced through the use of toxic language.