In response to the Meta Oversight Board's call for Public Comments, the Centre for Advanced Studies in Cyber Law and Artificial Intelligence (CASCA) , with insights from Dr. Ivneet Walia, Associate Professor of Law and Officiating Registrar at RGNUL and Research Consultant for the submission, submitted a detailed analysis regarding the moderation of political expression in Kenya. This submission addresses a specific instance where Meta's Hateful Conduct policy led to the removal of content using the term "Tugeges" on its platform.

The submission highlights that Meta faces significant challenges as its automated moderation systems struggle with linguistic nuances and local contexts in non-English languages, leading to ineffective detection and the amplification of hate speech in politically sensitive regions such as Kenya and India. This issue is further compounded by a disproportionate allocation of resources towards English content moderation and a notable lack of cultural competency among its content review teams , revealing a disconnect between platform policy and ground realities in the Global South.

CASCA proposes that Meta should engage with Natural Language Processing experts to utilize culturally-specific and open-source resources for improving moderation in diverse languages. They also recommend routine updates to algorithms, developed in consultation with civil societies, and made publicly available for transparency. Crucially, CASCA emphasizes the need for human intervention in content moderation, urging Meta to invest in local moderators who possess a deep understanding of the language, culture, and political contexts of their respective markets, and to ensure meaningful representation of minorities among its staff.