Ghana Sets Up Central AI Review System for All Public Sector Deployments

Ghana's government is establishing a centralised review mechanism for all artificial intelligence systems purchased by public sector organisations, signalling a strategic shift towards greater control over digital infrastructure and data security. Minister Samuel Nartey George announced the directive at the US Embassy AI Forum in Accra, framing the move as essential for protecting Ghana's digital sovereignty whilst building domestic technical expertise.

Under the new system, the Ministry of Communication, Digital Technology and Innovations will coordinate overall technical requirements, whilst the National Information Technology Agency (NITA) will handle the detailed technical assessment of proposed AI deployments across government agencies. This two-tier approach aims to create a consistent standard for how public institutions evaluate and adopt artificial intelligence tools, without disrupting existing public procurement frameworks.

What the New AI Review Will Cover

Government-acquired AI systems will be assessed against five key criteria: interoperability with existing government systems, data protection safeguards, cybersecurity standards, long-term sustainability, and local skills transfer. This comprehensive approach reflects concerns that acquiring AI systems without proper oversight can create vendor lock-in, expose sensitive government data, and drain resources into foreign technology providers rather than building Ghana's own technical workforce.

The directive represents a practical recognition that AI adoption in government requires more than conventional procurement rules. By making NITA the technical gatekeeper, the government aims to prevent individual ministries from acquiring incompatible or insecure systems independently, whilst ensuring that each AI investment strengthens Ghana's ability to manage its own digital infrastructure long-term.

Why it Matters for Ghana

This directive sits within President John Dramani Mahama's broader National Artificial Intelligence Strategy, which identifies eight key pillars for AI deployment. The government has already designated finance, agriculture, fisheries, education, health and communications as priority sectors for phased AI integration, recognising that these sectors could unlock significant productivity gains and improved service delivery for Ghanaians.

The centralised review system addresses a genuine risk: without coordinated oversight, government agencies could accumulate proprietary AI tools that only their original vendors can maintain, creating expensive dependencies and limiting Ghana's ability to audit how systems use public data. By insisting on data protection and local skills transfer, the directive attempts to ensure that public money spent on AI actually builds Ghana's domestic capacity rather than simply enriching foreign technology companies.

The government is backing this regulatory framework with substantial investments: a new AI Compute Centre is being developed, alongside upgrades to the National Data Centre. The One Million Coders Programme—which has already registered over 141,950 participants—signals intent to create a pipeline of Ghanaian software developers capable of building and maintaining AI systems locally. A proposed 10-year National AI Delivery Framework is expected to provide longer-term strategic direction for these efforts.

Implementation and Next Steps

The directive complements rather than replaces existing procurement laws, meaning government institutions will still follow standard tender and approval processes, but with an additional AI-specific technical review layer. This avoids creating bureaucratic redundancy whilst ensuring expert evaluation of AI-specific risks.

For public sector organisations, the change means that acquiring AI tools will require early engagement with NITA to assess technical fit. For technology vendors hoping to sell AI systems to Ghana's government, the directive signals that solutions must demonstrate strong data protection, compatibility with government infrastructure, and willingness to support local technical development.