Updated
Updated · Road to VR · May 6
GGWP proposes risk-based moderation for social VR safety
Updated
Updated · Road to VR · May 6

GGWP proposes risk-based moderation for social VR safety

4 articles · Updated · Road to VR · May 6
  • Across titles including Gorilla Tag and Animal Company, GGWP says under 1% of players drive about 28% of incidents, while risk-based sampling of 10% of sessions surfaces roughly 52%.
  • The company argues voice-first, immersive social VR makes blanket monitoring costly and intrusive, so developers should prioritize repeat offenders, session context and escalation workflows to cut harm more efficiently.
  • GGWP says targeted enforcement can improve deterrence, lower infrastructure costs and avoid perceptions of pervasive surveillance in social VR communities with limited moderation budgets.
If moderation AI only samples 10% of sessions, are new VR users being left unprotected?
As AI polices the metaverse, can VR be safe without becoming a total surveillance state?

Social VR Moderation Crisis: GGWP’s AI System Captures Over 50% of Harmful Incidents by Monitoring Just 10% of Voice Sessions

Overview

Social VR platforms face unique moderation challenges due to real-time voice and proximity-based interactions, making universal monitoring unsustainable and invasive. A small fraction of users cause most harm, so GGWP adopts a risk-based system that uses AI to calculate dynamic player reputation scores and focuses moderation on high-risk cases. This targeted approach monitors only a portion of voice sessions but captures the majority of harmful incidents, reducing costs and privacy concerns. GGWP’s system integrates cross-platform data, supports moderators with customizable tools, and emphasizes fairness and transparency. While scaling moderation remains complex, this model offers an effective, privacy-conscious path to safer, healthier social VR communities.

...