Abstract
Citizen-generated counter speech is a promising way to fight hate speech and promote peaceful, non-polarized discourse. However, there is a lack of large-scale longitudinal studies of its effectiveness for reducing hate speech. To this end, we perform an exploratory analysis of the effectiveness of counter speech using several different macro- and micro-level measures to analyze 131,366 political conversations that took place on German Twitter over four years. We report on the dynamic interactions of hate and counter speech over time and provide insights into whether, as in ‘classic’ bullying situations, organized efforts are more effective than independent individuals in steering online discourse. Taken together, our results build a multifaceted picture of the dynamics of hate and counter speech online. While we make no causal claims due to the complexity of discourse dynamics, our findings suggest that organized hate speech is associated with changes in public discourse and that counter speech—especially when organized—may help curb hateful rhetoric in online discourse.
Highlights
Hate speech is rampant on many online platforms and manifests in many different forms, e.g., insulting or intimidating, encouraging exclusion, segregation, and calls for violence, as well as spreading harmful stereotypes and disinformation about a group of individuals based on their race, ethnicity, gender, creed, religion, or political beliefs [1–10]
We want to solve problems together and act guided by mutual respect, love, and reason, We do not wage war but seek conversation. While both Reconquista Germanica (RG) and Reconquista Internet (RI) were active on several social media platforms, for our analysis, we focused on a sample of their Twitter presence, focused around several prominent German news organizations and public figures which enabled us to study the structure of the resulting conversations and the dynamic interplay between the two groups over time
2.2 Data For the analysis reported in this manuscript we performed two independent data collection phases: 1. Classifier Training Data Collection: We collected millions of Tweets originating from approximately 3700 known RG and RI members to train a classification system to identify hate and counter speech typical of these groups, as well as neutral speech not typical of either group
Summary
Hate speech is rampant on many online platforms and manifests in many different forms, e.g., insulting or intimidating, encouraging exclusion, segregation, and calls for violence, as well as spreading harmful stereotypes and disinformation about a group of individuals based on their race, ethnicity, gender, creed, religion, or political beliefs [1–10]. While it is widely accepted that hate speech is a growing problem on online platforms, what to do about it is a point of contention. One proposal is to more or less automatically detect and remove hateful content. This approach would have to overcome several challenges, including the nuanced and constantly evolving nature of hate speech, societal and legal norms about free speech, and the possibility of merely moving hate to other platforms rather than eliminating it [11]. Counter speech can take many forms, including providing facts, pointing to logical inconsistencies in hateful messages, attacking the perpetrators, supporting the victims, spreading neutral messages, or flooding a discussion with unrelated content [18–26]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
