UC Irvine

Games are more than media for consumption; they are increasingly valuable educational tools and social platforms connecting players around the globe. The known benefits aside, harmful conduct prevails in online play, severely impacting players, gaming communities, and the games industry at large. Players’ understanding of social safety and their coping strategies with harmful conduct are influenced by the design and implementation of the safety infrastructure in online games. Therefore, improving these mechanisms can help reduce harmful behavior, enable proactive coping strategies, and develop healthier gaming environments. Safety infrastructure is a broader concept commonly incorporating policies, technological systems, and trust and safety teams. This work focuses on the technological components that directly shape player experience through structuring and moderating social interactions.In three studies, (1) I assess parental controls and in-game tools of social safety and privacy, (2) examine how players perceive moderation systems, and (3) explore possible future avenues in AI-assisted moderation that move from common retributive approaches towards supportive interventions.I argue that existing retributive moderation systems are insufficient to cultivate thriving communities, and that currently missing proactive and supportive measures are better aligned with player needs and shared expectations.First, I show that retaliating against toxicity or leaving it unchallenged solidifies toxic norms; yet to a lesser extent, victimization and game experience also foster resilience and productive coping. Proactive tools that help players respond with constructive strategies rather than reciprocating harm are, therefore, pertinent. I use the Teens Online Safety Strategies (TOSS) framework to review parental tools and social safety controls in popular multiplayer online games and publishing platforms. I demonstrate that teens’ developmental needs are not well supported, and parental systems both infantilize them and leave them vulnerable. The available social safety and privacy controls are mostly restrictive, and mechanisms supporting active coping and impulse control are scarce, if they exist at all.The recent surge in utilizing AI invites the question of its potential for innovating moderation systems. Thus, I explore players’ attitudes toward real-time AI interventions aiming to reduce toxicity and foster prosocial behavior. I elucidate that the most opportune solutions seamlessly integrate AI-assisted mediation into gameplay and follow strong ethical considerations for transparency, privacy, as well as human agency and control.This dissertation delineates the current state of safety infrastructures and technological moderation in multiplayer online games and specifies viable future directions that may inspire a cascade of work exploring proactive supporting ways of moderation.