Solomon Asamoah on How AI Could Change Everyday Life in Ghana and Across Africa

AI could help Ghana plan roads, manage power, improve public services and widen access to useful information. It could also deepen inequality or waste scarce money. The outcome will depend on practical choices about people, infrastructure and funding.

The public debate about artificial intelligence often swings between excitement and fear. One side promises faster growth and better services. The other warns about lost jobs, false information and machines making decisions that people cannot challenge. Solomon Asamoah argues that Ghana and other African countries need a more useful discussion. AI is neither automatically good nor automatically harmful. Its effect on society will depend on where it is used, who benefits and whether the systems around it are strong enough to turn a new tool into a public gain.

That effect will reach far beyond technology companies. AI can influence how governments plan roads, how utilities find faults, how hospitals manage supplies and how farmers respond to weather. It can change what pupils learn, which skills employers value and how citizens receive public information. It may improve daily life in some places while making little difference in others. Where data is poor, electricity is unreliable or public bodies are weak, it can even add cost and confusion.

AI could make existing infrastructure work better

The first opportunity is not a futuristic city built from scratch. It is getting more value from infrastructure that already exists. A power company can use AI to spot unusual demand, predict equipment failure and reduce losses. Road agencies can study traffic and maintenance records to decide which repairs are most urgent. Water utilities can search for leaks by comparing pressure, flow and billing data. Ports can plan the movement of goods more smoothly, while public transport operators can adjust services to actual passenger demand.

These uses will not remove the need for engineers, maintenance teams or new investment. A damaged transformer still has to be replaced, and a road still has to be built. AI can help people decide earlier and use limited budgets more carefully. The benefit is therefore less about replacing physical infrastructure than improving the way it is planned and operated.

There is a risk, however, that attention shifts from basic needs to attractive digital projects. A city does not become smarter because it buys software while drains remain blocked or traffic signals do not work. Public bodies should begin with a clear problem and ask whether better data and AI can help. If the underlying service is poorly organised, adding an automated layer may simply make failure harder to understand.

Public services could become more responsive

Health, education and agriculture show how the technology could reach ordinary households. Clinics can use simple forecasting tools to estimate demand for medicines. Health workers can receive support when reviewing scans or organising patient records, provided that trained professionals remain responsible for care. Teachers can prepare materials for pupils at different levels, while education authorities can identify where attendance or learning is falling behind. Farmers can receive local advice on pests, rainfall and prices.

The same tools can also exclude people. A service delivered only through a smartphone will miss citizens without affordable data, reliable coverage or confidence online. A health system that uses incomplete records may give worse advice to rural patients. An automated benefits system can reject a valid claim without explaining why. The measure of success should therefore be whether a service becomes easier to reach and more reliable, not simply whether an AI system has been installed.

For Solomon Asamoah this is an investment test as much as a technology test. Spending on software must be matched by spending on reliable records, staff training, connectivity and ways for the public to report errors. Otherwise, governments may purchase impressive systems that cannot work properly in the institutions expected to use them.

Language is economic infrastructure

A system trained largely on English-language material may appear capable in Accra while remaining unreliable for a trader, farmer or patient communicating in Twi, Ewe, Ga, Dagbani or another local language. This is not simply a cultural concern. Language determines who can use a service, what knowledge can be represented and whether errors can be detected by the people most affected.

Local-language material should therefore be treated as part of the region's essential digital foundation. Universities, broadcasters, publishers, traditional authorities and technology firms can help create it, but contributors and communities should know how their words and voices will be used. Governments can support open resources where there is a clear public benefit while protecting private, sacred or commercially sensitive material.

Smaller systems built for a clear local task may be more useful than the largest general tools. A service designed to transcribe a limited set of clinical terms or answer questions about one agricultural programme can be cheaper to test and easier to correct. African progress in AI is likely to come from careful local use, not from trying to copy every expensive model built elsewhere.

Work will change, but not all at once

Claims that AI will quickly remove millions of African jobs overlook the way people actually work across the continent. An IMF analysis published in 2026 estimated that about four-fifths of jobs in sub-Saharan Africa currently have limited exposure to AI. Among the jobs likely to be affected, it placed about 13 per cent of total employment at significant risk of replacement and almost 10 per cent in roles more likely to be helped by the technology.

Change is likely to reach office, finance, media and professional work sooner than many manual and informal jobs. Even there, the more common effect may be a change in tasks. Accountants may review automated summaries. Journalists may use faster research tools but carry greater responsibility for checking false material. Civil servants may process routine requests more quickly while handling difficult cases themselves.

Schools and employers still need to prepare. Training should not be limited to programmers. Teachers, nurses, managers and public officials need to know how to check an answer, protect personal information and decide when a machine should not be trusted. If productivity rises, employers and governments will also face a social choice about whether the gains support better services and wages or only reduce staff numbers.

Funding should follow public value

Ghana cannot fund every promising idea, and it should not try. Early public spending should focus on shared foundations that many users need, including reliable power, affordable broadband, useful public data, research capacity and digital skills. Individual applications should then compete for support based on the problem they solve and the evidence that they work.

Funding can come from several sources. Government budgets can support public records and core services. Development banks can provide longer-term finance for networks, training and shared computing facilities. Private companies can invest where there is a clear customer and revenue stream. Universities and regional bodies can pool demand for research tools that would be too expensive for one institution. Partnerships can help, but contracts must state who owns the data, who pays when costs rise and what happens if the technology fails.

Solomon Asamoah brings more than three decades of investment experience to these choices. His senior roles have included the International Finance Corporation, the Development Bank of Southern Africa, the Africa Finance Corporation and the African Development Bank. He later served as the first substantive chief executive of the Ghana Infrastructure Investment Fund where his work involved turning national priorities into projects that could be assessed, funded and managed over time.

The outcome is still open

AI will probably be positive in some parts of African life, negative in others and barely noticed in many more. That is not indecision; it is the normal result of a tool meeting different institutions and communities. Ghana can improve the balance by testing systems on a small scale, publishing what worked, stopping what did not and giving citizens a clear way to challenge decisions that affect them.

The African Union's Continental AI Strategy, endorsed in Accra in July 2024, offers a useful direction by linking innovation with skills, data, inclusion and responsible use. The next step is to turn those aims into choices that people can see in schools, clinics, farms, utilities and local government.

The real measure will not be the number of AI announcements. It will be whether infrastructure lasts longer, public money goes further, workers gain useful skills and services reach people who were previously overlooked. Solomon Asamoah's view is that Ghana and Africa can benefit from AI without accepting every use of it. Clear priorities, careful funding and institutions that remain answerable to the public will decide whether the technology strengthens society or simply adds another layer of cost.

Key factual references used in this draft:

• African Union, Continental Artificial Intelligence Strategy (2024)

• IMF, Unlocking the Potential: AI in Sub-Saharan Africa (2026)

• World Bank, Digital Progress and Trends Report 2025: Strengthening AI Foundations