GHANA ORDERS CENTRAL REVIEW OF ALL PUBLIC SECTOR AI PURCHASES

Ghana's government has introduced a centralized technical review for artificial intelligence systems purchased by public institutions, creating a new layer of oversight as state agencies increasingly explore AI for administration, public services and data analysis.

Minister for Communication, Digital Technology and Innovations Samuel Nartey George announced the directive in Accra, saying the government wants to prevent public institutions from independently purchasing AI systems that may be incompatible, insecure, unsustainable or overly dependent on a single foreign technology provider.

Under the new arrangement, the Ministry will coordinate the technical requirements for proposed AI deployments.

The National Information Technology Agency, known as NITA, will lead technical scoping and assessment.

The government says the process is designed to strengthen oversight before public money is committed to artificial intelligence systems.

AI PURCHASES WILL FACE TECHNICAL SCRUTINY

The reviews will examine several major areas.

One is interoperability.

Government systems increasingly need to exchange information across ministries, departments and agencies.

If one institution purchases an AI platform that cannot communicate with existing government infrastructure, it can create expensive duplication and make future integration difficult.

The new review process is intended to identify those problems before contracts are signed.

Cybersecurity will be another major area of assessment.

AI systems can process large quantities of sensitive information, including government records and potentially personal data belonging to citizens.

Poorly secured platforms could create vulnerabilities that expose state systems to unauthorized access, data theft or manipulation.

DATA PROTECTION WILL BE CENTRAL

The government also says proposed systems will be assessed for compliance with data protection requirements.

That issue is particularly important because many modern AI systems depend on large datasets.

Public institutions can hold highly sensitive information involving health, education, taxes, identity, employment and social services.

If those datasets are transferred into external AI platforms without adequate safeguards, questions arise over where the data is stored, who can access it and whether it can be used for purposes beyond the original government function.

The centralized review is intended to make those questions part of the procurement process rather than an afterthought.

GOVERNMENT WANTS TO AVOID VENDOR LOCK IN

Another major concern is dependence on individual technology companies.

Vendor lock in occurs when an organization becomes so dependent on one supplier's technology that changing providers becomes difficult or expensive.

This can happen when systems use proprietary data formats, closed software architecture or specialized tools available only from one vendor.

For government institutions, vendor lock in can create long term financial and strategic risks.

A ministry may initially purchase a system at an affordable price but later face high upgrade, maintenance or licensing costs because switching to another provider would require replacing large parts of the infrastructure.

The new AI review will therefore consider whether proposed systems allow sufficient flexibility and interoperability.

LOCAL SKILLS TRANSFER WILL ALSO MATTER

The government says public AI acquisitions will also be assessed according to how much knowledge and technical capacity they leave inside Ghana.

This means contracts may increasingly be judged not only by the technology being purchased but by whether Ghanaian professionals receive the training necessary to operate, maintain and improve those systems.

Skills transfer has become an important issue as African governments sign large technology agreements with international companies.

A country can spend millions of dollars on sophisticated digital infrastructure while remaining dependent on foreign engineers if domestic workers are not trained.

Ghana says it wants AI investments to strengthen local technical capacity rather than simply create long term dependence on external expertise.

SUSTAINABILITY IS PART OF THE REVIEW

Sustainability will also form part of the technical assessment.

Artificial intelligence systems can require significant computing resources, data center capacity, electricity and long term maintenance.

Government agencies may be attracted by the immediate capabilities offered by a new system without fully considering its operational cost over several years.

The centralized process is intended to examine whether institutions can realistically maintain the technology after the initial acquisition.

That includes assessing infrastructure requirements, recurring software costs, technical staffing and long term support.

THE DIRECTIVE DOES NOT REPLACE PROCUREMENT LAW

The government has emphasized that the new mechanism will not replace Ghana's existing procurement procedures.

Public institutions will still be required to follow the legal processes governing the expenditure of public funds.

The technical review instead adds another level of scrutiny.

In practical terms, an institution could satisfy ordinary procurement requirements but still face questions about whether the AI system it wants to purchase is technically suitable for government use.

The aim is to separate two questions.

Was the procurement process legally conducted?

And is the technology itself appropriate, secure and sustainable?

The new framework attempts to address both.

NITA ALREADY HAS A TECHNICAL CLEARANCE ROLE

The directive also builds on a broader government effort to centralize technology oversight.

Earlier in 2026, NITA institutionalized a Technical Clearance Framework for public sector information technology investments.

The system is intended to identify duplicate or nonconforming digital projects before public funds are committed.

Samuel Nartey George said in September that six government institutions and projects had already received technical clearance under the framework.

NITA had also provided technical assurance to several public institutions and reviewed digital health and business registration systems.

The new AI review therefore appears to extend an existing government IT oversight model into a more specialized and rapidly developing area.

WHY AI NEEDS DIFFERENT OVERSIGHT

Artificial intelligence creates procurement challenges that differ from traditional government software.

A normal database or administrative system generally performs predictable tasks.

AI systems can generate new content, classify data, make recommendations and identify patterns based on large datasets.

Some systems can also change in performance as models are updated.

That creates questions that ordinary procurement rules may not fully address.

Who is responsible when an AI system produces an inaccurate recommendation?

How can government audit the decisions of a system developed by an external company?

Can sensitive government data be used to train a vendor's model?

What happens when the vendor changes its pricing or technical terms?

These are increasingly important questions for governments around the world.

AI IN GOVERNMENT CAN BRING REAL BENEFITS

The tighter review process does not mean Ghana is turning away from artificial intelligence.

The government is actively promoting AI adoption.

AI can potentially help public institutions process documents faster, detect fraud, improve agricultural forecasting, analyze health information and reduce administrative delays.

In education, AI can support personalized learning and automate routine tasks.

In agriculture, it can analyze weather, crop and soil information.

In health care, AI can assist with diagnostics and disease surveillance.

Government agencies can also use AI to identify patterns across large datasets that would be difficult for human analysts to process manually.

The challenge is gaining those benefits without creating new security, privacy or financial risks.

GHANA'S NATIONAL AI STRATEGY

The new procurement directive forms part of Ghana's broader National Artificial Intelligence Strategy.

The strategy identifies several areas where AI could support economic development and public services.

These include finance, agriculture, fisheries, education, health and communications.

The government has said implementation will be phased rather than allowing uncontrolled adoption across institutions.

That approach reflects a broader international trend.

Governments increasingly want to benefit from AI while establishing rules for accountability, safety and data governance.

DIGITAL SOVEREIGNTY IS BECOMING A POLICY PRIORITY

One of the ideas behind Ghana's AI policy is digital sovereignty.

The concept refers broadly to a country's ability to maintain meaningful control over its data, digital infrastructure and technological capabilities.

For developing countries, this can be difficult.

Much of the world's major cloud infrastructure, AI software and advanced computing capacity is controlled by companies based outside Africa.

That means governments can become dependent on foreign platforms for essential digital services.

Ghana's emphasis on interoperability, local skills and reduced vendor dependence suggests the government wants to limit that exposure.

Digital sovereignty does not necessarily mean rejecting foreign technology.

It means attempting to ensure that foreign partnerships do not leave the country without control over critical systems.

GHANA IS ALSO INVESTING IN COMPUTING INFRASTRUCTURE

The government is pursuing additional investments in national digital infrastructure alongside the new review regime.

Plans include an AI Compute Center and upgrades to the National Data Center.

High performance computing infrastructure is increasingly important because advanced AI models require significant processing capacity.

Countries without local computing infrastructure can become dependent on foreign cloud providers.

That creates additional costs and can raise questions about where sensitive national data is processed.

Building local capacity could allow Ghanaian institutions, universities and technology companies to run more advanced AI systems domestically.

NATIONAL DATA EXCHANGE HUB ALSO UNDER DEVELOPMENT

The government is also advancing a National Data Exchange Hub.

The project is intended to allow public institutions to exchange information more efficiently.

It forms part of the World Bank backed Ghana Digital Acceleration Project.

Interoperability is a key part of such infrastructure.

If ministries and agencies purchase incompatible systems, a national data exchange platform becomes much harder to operate.

The new AI technical review could therefore help align future acquisitions with Ghana's wider digital architecture.

ONE MILLION CODERS PROGRAM FITS INTO THE SAME STRATEGY

The government is also linking artificial intelligence policy with workforce development.

The One Million Coders Program is intended to expand digital skills among Ghanaians.

Government says more than 141,000 registrations have already been recorded.

The logic is straightforward.

Ghana cannot build a sustainable AI ecosystem if it imports both the technology and the people required to operate it.

Training local developers, data specialists, cybersecurity professionals and AI engineers is therefore central to reducing long term dependence on foreign expertise.

CYBERSECURITY COULD BECOME THE BIGGEST RISK

As government services become more digital, the consequences of cyberattacks become more serious.

A compromised entertainment website may inconvenience users.

A compromised government AI system could potentially expose personal records, disrupt public services or produce false information used in official decision making.

AI can also create new attack opportunities.

Adversaries may attempt to manipulate training data, exploit weaknesses in software or extract sensitive information from models.

Technical reviews before deployment cannot eliminate those threats.

But they can reduce the risk of government agencies purchasing systems without adequate security standards.

PUBLIC MONEY IS ALSO AT STAKE

AI has become one of the world's most heavily marketed technologies.

Technology companies increasingly promote AI products to both businesses and governments.

That creates a risk that institutions purchase systems because AI is fashionable rather than because the technology solves a clearly defined problem.

Central review could help government determine whether an AI system is actually necessary.

In some situations, conventional software may perform the required function more cheaply and reliably.

Public institutions therefore need to evaluate not simply whether a product uses AI but whether it delivers measurable value.

CENTRALIZATION CAN ALSO CREATE CHALLENGES

The policy will also need to avoid creating excessive bureaucracy.

Central review can improve standards, but it can slow innovation if institutions face long approval delays.

NITA and the Communications Ministry will therefore need sufficient technical staff to review potentially complex AI systems quickly.

The process will also need transparent standards.

Public institutions and technology vendors should understand what requirements must be met before a project can proceed.

If decisions are inconsistent or unclear, the review system itself could become a source of procurement delays.

TRANSPARENCY WILL BE IMPORTANT

Another question is how much information the government will make public.

AI procurement can involve large contracts and systems capable of affecting significant numbers of citizens.

Transparency around the purpose, cost and capabilities of such systems can strengthen public trust.

Government could also consider publishing information about which systems receive technical approval and the standards used to evaluate them.

That would allow citizens, civil society and technology experts to better understand how AI is being introduced into public administration.

GHANA IS MOVING EARLY

Many African governments are still developing national frameworks for artificial intelligence.

Ghana's decision to create a dedicated review process for public sector AI acquisitions places it among governments attempting to establish oversight before AI adoption becomes widespread.

That timing could be important.

It is easier to establish common standards before dozens of institutions purchase incompatible systems than to correct those problems afterward.

The financial cost of replacing poorly designed technology can be substantial.

The security cost can be even greater.

WHAT HAPPENS NEXT

The practical success of the directive will depend on implementation.

Government institutions planning AI projects will now need to engage the Communications Ministry and NITA before moving ahead.

NITA will need to conduct technical assessments covering cybersecurity, interoperability, data governance and long term sustainability.

The government will also have to determine how the centralized review interacts with public procurement timelines.

Technology suppliers may face stronger requirements to demonstrate how their systems protect data and integrate with existing infrastructure.

They may also face pressure to provide more training for Ghanaian professionals.

The policy represents a shift from simply encouraging AI adoption to actively managing how it enters government.

That distinction matters.

Artificial intelligence can improve public services, reduce administrative costs and help institutions make better use of data.

But poorly managed adoption can create expensive systems that do not communicate with each other, expose sensitive information or lock governments into long term dependence on individual vendors.

Ghana's new approach is an attempt to capture the benefits of AI while reducing those risks before public money is spent.