Abstract Online communities are vital civic infrastructure, yet the moderators who sustain them remain poorly supported by existing tools. This article introduces a framework for translating ethnographic insight into AI products that effectively augment human judgement. We understand community moderation as collective intelligence—emerging from the interplay of moderators, members, rules, and technical systems. Our approach, which we describe in this case, combined three methodological innovations: 1. applying an ecological lens to study moderation as a system rather than isolated acts; 2. developing ethnographic exercises to surface tacit knowledge and judgment; and 3. convening a co‐design workshop where moderators and product teams prototyped together. These methods made visible how community moderation relies on shared purpose, contextual reasoning, and relational work that current tools overlook.
Abonizio et al. (Sat,) studied this question.