Malloum: An SMS-Based Generative AI Tutoring System for Contextualized Learning in Low-Connectivity Environments

Abstract

Limited internet connectivity continues to constrain access to digital learning resources across many low- and middle-income countries, particularly in Sub-Saharan Africa. While generative artificial intelligence has demonstrated strong potential for personalized tutoring, most existing systems assume reliable broadband access and smartphone availability. This paper presents Malloum, an SMS-based generative AI tutoring system designed to deliver contextualized, curriculum-aligned pedagogical support in low-connectivity environments. Malloum enables students to submit academic questions via standard SMS to a dedicated short code. The system integrates a backend orchestration layer with a hierarchical prompt architecture leveraging a state-of-the-art generative language model. A multi-layered prompting strategy combines a system-level pedagogical role definition, contextual curriculum injection aligned with the Cameroonian national syllabus, and user-level query input. An educational database (including validated exam banks, reference materials, and structured student profiles) enhances response accuracy, personalization, and pedagogical alignment. To address SMS-specific constraints, the platform implements structured micro-explanations, intelligent compression algorithms, automatic message concatenation, and adaptive bilingual output (French/English). A matched-pair cluster-assigned field study conducted with 1,000 secondary school students across 20 schools in the Far North Region of Cameroon using a pre-test/post-test design with a comparison group shows that the mean composite score of the Malloum group rose from 51.7% to 72.4%—an absolute gain of 20.7 percentage points, corresponding to a 40% relative improvement over baseline—compared with a gain of 4.3 percentage points (8.3% relative) in the control group, yielding an adjusted between-group difference of 16.1 percentage points (95% CI: 13.9–18.3, p < 0.001). Malloum provides a scalable, cost-aware, and culturally contextualized framework for AI-driven education, offering a practical pathway to narrowing the digital divide in secondary education systems.