Algorithmic Moderation and Conflict Narratives in Nigeria: How Platform Decisions Shape Amplification, Suppression, and Peacebuilding Outcomes

Algorithmic content moderation determines which conflict narratives circulate, and the scholarship examining it has concentrated on high-resource language environments in consolidated democracies. This article argues that the concentration produces analytical distortion rather than merely a coverage gap, because moderation systems in high-resource settings operate under conditions that do not obtain elsewhere. Employing a convergent mixed-methods design, the study combines computational analysis of 5.26 million posts from X and Facebook across three politically sensitive event windows with 15 semi-structured interviews involving platform policy and trust-and-safety professionals, examining content in English, Hausa, Yoruba, and Igbo. The findings identify systematic disparity, with English-language content approximately 1.9 times more likely to be moderated than Hausa-language content within comparable thematic categories, and with political figures subject to substantially higher enforcement than ordinary users. The direction of the disparity is analytically significant, since lower-resource language content is under-enforced rather than over-enforced, which indicates classifier absence rather than classifier error. Practitioner accounts attribute the pattern to training data scarcity, policy framing developed for other contexts, and review capacity allocated by content volume rather than risk exposure. The article advances the claim that moderation disparity in conflict-affected settings constitutes a peacebuilding deficiency rather than only a technical one, and it qualifies the transparency remedy that the literature conventionally recommends by examining the Nigerian regulatory record, where disaggregated reporting is already a legal obligation, where the 2024 compliance assessment found incomplete and non-comparable submission, and where platform rule enforcement was paused in July 2026 pending a unified digital policy.