Somalia’s AI gap is bigger than connectivity

Digital access is only the first dependency

Somalia’s energetic telecoms and mobile-money sectors do not, by themselves, make the country ready to benefit broadly from AI. A peer-reviewed systematic review concludes that the binding constraints also include fragmented governance, weak data ecosystems, higher-education programmes poorly aligned with employers and no coordinated national AI or digital-skills policy. For workers, that means buying tools before building institutions could deepen exclusion rather than improve productivity.

The authors searched Scopus, Web of Science, Google Scholar, PubMed and ERIC in July 2026, supplemented by publisher platforms and reference tracking. They identified 345 records, removed 35 duplicates, screened 310 titles and abstracts, reviewed 71 full texts and included 31 studies in a narrative synthesis. Two reviewers screened and appraised studies, with disagreements resolved by discussion and a third reviewer if needed.[1]

The evidence points to four linked capabilities

The proposed framework has four pillars: infrastructure and data governance; institutional readiness; human capital and skills; and labour-market integration. The sequence matters. Connectivity and data access enable adoption, but organisations need accountable leadership and workers need enough technical and domain knowledge to use systems critically. Training has limited value if recruitment and curricula remain disconnected from real jobs.

That diagnosis fits Somalia’s unusual structure: private digital services expanded faster than formal public capacity. The review therefore proposes public–private interoperability rules, coordinated government strategy, curriculum reform, sustained instructor development and university–employer alignment. It also recommends starting sector pilots where digital capability already exists, particularly telecommunications and mobile financial services, rather than assuming every public service is equally ready.

The paper is explicit that these are analytical proposals, not a validated implementation blueprint. Its included literature spans different designs and outcomes, preventing meta-analysis. Searches were conducted only in English, and Somalia-specific empirical research was scarce—especially in health, agriculture and public administration. Evidence borrowed from comparable African or fragile settings may not transfer cleanly to Somalia’s institutions and political economy.[1]

A strategy still needs measurable tests

For a graduate or public-service worker, the practical question is not whether an AI strategy exists, but whether it produces accessible training, fair recruitment, safe data practices and better work. The roadmap would need denominators: who receives training, who completes it, who gains relevant employment, which regions remain excluded, and whether productivity improvements survive after donor or pilot support ends. It would also need comparison groups or phased roll-outs capable of separating programme effects from wider economic change.

The evidence check also includes interests. The authors reported no external funding, but one was employed by education-technology company Dugsiiye, and the group acknowledged Somali National University and Somali AI Network. Stronger evidence would combine representative workforce surveys, employer vacancy and skills data, connectivity and affordability measures, longitudinal programme evaluation and independent audits of public–private data governance. The review usefully reframes readiness as capability-building; it does not prove that its four-pillar sequence will work.[1]