RightsCon 2025 Hybrid Roundtable: Exploring Abuse Detection Challenges and Contextual Biases in AI-Driven Hate Speech Moderation

RightsCon 2025 Hybrid Roundtable: Exploring Abuse Detection Challenges and Contextual Biases in AI-Driven Hate Speech Moderation 

25 February

AI-driven hate speech moderation, shaped largely by Western-centric frameworks, reinforces linguistic and geopolitical exclusions, failing to capture the sociocultural specificities of non-dominant languages and politically vulnerable regions. As a result, these systems often silence critical voices while allowing harmful content to persist. The discussion will explore how opaque algorithmic decision-making, data gaps, and governance structures exacerbate rather than mitigate harm, further entrenching digital inequalities.

Moderated by Sadaf Wani
Speakers: Yasmin Curzi (Brazil), Wangari Njathi (United States of America), Rebeccah Wambui (Kenya), Caitlin Kraft Buchman (Switzerland).