Disinformation and Hate Speech as Intertwined Online Harms: A Large-Scale Study Using Fact-Checking Data

Discourse on social media is produced and governed through distributed cooperative work, yet information operations exploit these sociotechnical systems by combining epistemic disruption (disinformation) with antagonistic boundary enforcement (hate speech). This study examines when and how disinformation and hate speech become intertwined in online discourse, conceptualising echo chambers as escalation and collective alignment processes and identity work as boundary-enforcing othering. Drawing on large social media fact-checking datasets, we develop a scalable method that annotates hate speech with automated detection, validated against human annotations and relates it to disinformation ratings across topical domains. We find a statistically reliable but modest overall coupling between disinformation and hate speech that is strongly topic-contingent: Associations are elevated in discourses on race and ethnicity, while weaker patterns appear in domains such as Covid-19 and entertainment. Analyses of targeting suggest that boundary enforcement can either broaden to additional out-groups or focus on topic-salient targets, depending on topical context and category definitions. We discuss implications for cooperative platform governance, arguing that cross-harm coupling should function as a topic-sensitive triage signal to support coordination between moderation and fact-checking workflows rather than as an automated enforcement rule.