What happened
Three ministers leading national digital and AI strategies set out, in concrete terms, why open source is central to development, not as a cost-saving afterthought, but as the mechanism to inspect, adapt, secure and govern the technology a country runs on. Across very different contexts, the same argument recurred: proprietary, vendor-locked systems trap public budgets in perpetual licensing and rarely fit local needs, while open source lets a country build on what already exists rather than start from scratch. The tension they kept returning to was trust: open source is only viable if it is maintained, supported, and demonstrably reliable.
Key points
- Morocco: contextualization and governance. Minister El Fallah Seghrouchni noted that open source for AI is harder than for software: releasing code isn’t enough without data, weights and the means to fine-tune. Morocco is building “data factories” and a “forge” of shared sources, adapting models to under-served languages such as Amazigh, and standing up a network of regionally-specialized institutes backed by more than a gigawatt of compute. Her caution: open source can improve the explainability of AI, but only if a country has the talent to actually read and understand the code.
- Jamaica: capability, not dependency. Minister Marks described open source as “fundamentally a capability issue, not just a technological one,” backed by a National AI Task Force, a UNESCO readiness assessment, a national AI lab, a new data-exchange platform, and a national ID, plus hackathons that grew from 150 to 750 young participants. The goal: move “from dependency to active participation and contribution.”
- Sierra Leone: open source first, and shared infrastructure. Minister Bah cast open source as a way to “change the economics of digital transformation” and escape the licensing “trap.” Sierra Leone is a founding member of the Digital Public Goods Alliance, has adopted an open-source-first policy, and is pushing a regional “data embassy” framework so West African states can share compute and data infrastructure rather than each building their own, a chance, she argued, to “leapfrog decades.”
Recurring themes
- Fragmentation is the enemy. All three flagged the waste of neighbouring countries building the same solutions in parallel, and the case for regional pooling of talent, data, and infrastructure.
- Trust is the barrier, not cost. The fear that open source means “no maintenance, nobody to call” (and that “free means cheap, cheap means bad”) is what slows adoption. Local talent contributing to (not just consuming) open source is how that narrative changes.
- The private sector was missing. Sierra Leone’s minister pointedly wished a private-sector voice had been on the panel, arguing that representation in AI (whose stories and whose data shape the models) is a shared responsibility, not governments’ alone.
Why it matters for the SDGs
This was the day’s clearest line to the Goals. Each minister tied open-source AI to concrete public-service outcomes (health, education, agriculture, social protection, and cheaper, more accountable government) and to SDG 9 (infrastructure and innovation), SDG 10 (reducing the AI divide), and SDG 17 (partnerships, shared infrastructure, and South-South cooperation). The shared conviction: the point of AI is not AI, but whether a country can own and improve the systems that serve its citizens.
Watch & read
- UN Web TV, recording of the AI Day (23 June 2026).
- Digital Public Goods Alliance, referenced by Sierra Leone as a founding member.
- Full AI Day recap, the day’s other sessions.
Ministers’ names and portfolios reflect roles at the time of the event; quotations are lightly edited from an automated transcript and should be read as close paraphrase.