Introducing Gamified Reputation Metrics for Self-Regulating User Behavior in Social Media

Online hostility in social media, including hate speech, trolling, and cyberbullying, can seriously harm people's mental and physical health, and yet it is often addressed after harm has already occurred. While prior research focused on detecting such behaviors, this study proposes and evaluates a gamified, user-centered approach to prevent and regulate hostile interactions while preserving user agency. To this end, we have designed and developed an innovative gamified algorithm that combines elements such as Corrected Hostility Level (CHL), Politeness Points, green cards, apology and appeal mechanisms, user-selected politeness thresholds, Politeness Badges, and a Positive Reputation Score (PRS). The proposed system enables users to restrict unwanted interactions without requiring the platform to read private messages or impose uniform moderation rules. The proposed methods were implemented in Gegli Social Network, and the evaluation analyzed 36,367 reports concerning 16,167 reported users and 97,818 visits to the politeness panel by 1,478 users. Moreover, the PRS analysis used 132,588 testimonial records and identified 24,151 ranked receivers, showing a strongly right-skewed reputation distribution in which early badge levels were widely reachable and higher badges remained rare. The study results suggest that the gamified mechanisms were associated with reduced hostile interactions, motivated users to apologize to each other for their bad behavior, and personalized their communication by democratically restricting their communication to decide with whom they wanted to communicate, while mostly selecting the most open threshold. This is to encourage safer and more prosocial online behavior. Gender differences in the use of the proposed method were also studied.