Before the Machines, the Rules: A Policy Critique of Ghana’s AI Spending, the Proposed $250 Million AI Centre, and the Publican Customs Rollout
Ghana’s AI turn is now real. The strategy has moved into Cabinet space, AI is already being deployed in customs, and government is clearly trying to connect digital skills, data governance and public sector AI. The problem is not the ambition. The problem is that the public rules, financing detail and accountability architecture still trail the pace of the announcements. That gap is what makes the proposed $250 million AI centre and the Publican customs rollout policy-significant and governance-sensitive.
Executive summary
The public record now supports a clear conclusion that Ghana is no longer merely speaking about artificial intelligence in broad developmental terms. The Ministry of Communication, Digital Technology and Innovations has publicly tied AI to a national strategy, skills development, data protection reform, an emerging technologies bill and a series of institutional partnerships. The 2026 Budget Speech also goes beyond rhetoric by requesting approval for an AI-powered trade data analytics system to support customs revenue mobilisation and by allocating GHS 100 million for the National Coders Programme. In other words, AI has entered Ghana’s policy machinery.
At the same time, the most politically striking AI announcement in the current cycle, the proposed $250 million AI centre, is not yet publicly specified with the level of detail one would expect for a project of that scale. The public materials reviewed do not yet show whether the centre is a straightforward budgetary expenditure, a public-private partnership, a blended-finance project, or a donor-backed initiative. Nor do they conclusively establish whether it is simply a re-scoped version of the earlier National Artificial Intelligence Innovation Centre referenced in the Ministry’s Q2 2025 briefing, or whether it is a new flagship project sitting alongside that earlier initiative.
The Publican customs project makes the governance issue even sharper. Customs is one of the strongest public-sector use cases for AI because the data are structured, the abuse patterns are known, and the revenue stakes are high. But the memorandum available through public disclosure which was reviewed for this piece, read on its face, suggests an operational posture in which AI outputs may become the effective floor for valuation decisions, with a centralised appeals route and disciplinary pressure on officers who do not use the tool as directed. That is where a legitimate revenue-assurance tool can drift into a due process and accountability problem.