This study aims to evaluate a community-based artificial intelligence (AI) and data science practicum designed to strengthen the knowledge, attitudes and applied competencies of health information science students in Tanzania. The practicum responds to persistent gaps in AI preparedness within health curricula in low-resource settings, where infrastructural constraints, rural service delivery and linguistic diversity shape both learning and practice.
A programme-evaluation design was used, using pre- and post-intervention assessments to examine changes in students’ knowledge, attitudes and practical skills. The practicum was informed by experiential learning and diffusion-of-innovation theories and delivered through short conceptual lectures, bilingual instructional materials, hands-on analytics exercises and community-linked data projects using offline-capable tools. Quantitative outcomes from 27 practicum participants were complemented by qualitative reflections to assess learning processes and contextual fit. A contemporaneous non-trained cohort was described for background comparison but not used for causal inference.
Participants demonstrated large and statistically significant gains across all learning domains following the practicum. Knowledge of AI and data-science concepts increased substantially, attitudes shifted from neutral to strongly positive and practical competence improved in tasks such as data cleaning, basic modelling and applied analytics. Learning gains were most pronounced where activities directly reflected Tanzanian public health priorities and operational constraints. The integration of locally relevant data sets, bilingual delivery and offline workflows proved central to overcoming digital and linguistic barriers.
The findings show that complex AI and data-science concepts can be translated into usable competence through short, context-aware, community-anchored training. The practicum offers a scalable and resource-conscious model for integrating AI education into health information programmes in low-resource settings, with relevance for educators, curriculum designers and health-sector policymakers.
Rather than emphasising between-group comparisons, this study advances understanding of how contextual tailoring operationalises experiential learning in AI education. It contributes empirical evidence from a low-resource African setting and provides openly available learning materials, data sets and assessment tools to support replication and adaptation in similar environments.
