In this thesis, I explore the challenge of enabling rich governance mechanisms in end-to-end encrypted (E2EE) messaging platforms while preserving the privacy and security benefits of E2EE. It introduces the concept of private, hierarchical governance which allows user communities to self-moderate according to customizable policies while enabling escalation to platform-level moderation when needed. A modular system design is presented consisting of a messaging layer that provides ordered delivery of governance messages and a governance layer that executes expressive policies on the client side. The system is implemented in a prototype called MlsGov and an experimental evaluation demonstrates its practicality and ability to enforce complex governance procedures like user voting. I also examine how recent advancements in machine learning could further enhance automated content moderation capabilities, even in E2EE settings. Overall, this work demonstrates that private, hierarchical governance is a viable approach for balancing user privacy with content moderation needs in E2EE messaging platforms.