Intersectional Moderation: Designing AI and Human Systems for Marginalized Communities

Thach et al. (2024) critique mainstream moderation as insufficiently responsive to marginalized groups, while Balendra (2025) and Wu & Semaan (2023) highlight the failures of AI and decentralized models in protecting vulnerable users (e.g., trans communities, racial minorities). This research would co-design, with affected communities, hybrid moderation systems that blend AI, community moderators, and participatory feedback, explicitly incorporating intersectional needs (e.g., transphobia, racial microaggressions, covert discrimination). Rather than imposing abstract “neutrality,” the framework would recognize and foreground lived experience as a core input to governance. The project would measure outcomes on safety, trust, and user empowerment, offering a template for platforms seeking to address the unique governance challenges faced by marginalized groups—pushing the field toward more inclusive, equitable digital societies.

References:

1. Meta’s AI moderation and free speech: Ongoing challenges in the Global South. Soorya Balendra (2025). Law and Governance.
2. Trans-centered moderation: Trans technology creators and centering transness in platform and community governance. Hibby Thach, Samuel Mayworm, Michaelanne Thomas, Oliver L. Haimson (2024). Conference on Fairness, Accountability and Transparency.
3. "How Do You Quantify How Racist Something Is?": Color-Blind Moderation in Decentralized Governance. Qunfang Wu, Bryan Semaan (2023). Proc. ACM Hum. Comput. Interact..