The Honourable Sam Dzata George‘s launch of the National AI Strategy earlier today appears as a leap into the Fourth Industrial Revolution. The unveiling of the demo Aku AI chatbot, was supposed to be the proof. Instead, it was a mirror.
The Aku AI demo reflected not Ghana’s ambition, but the sad reality of our government’s backwardness: a state-led attempt to achieve in 2026 what the private tech sector practitioners mastered years ago.
Yet within this government backwardness lie real possibilities, and even greater risks, that will define whether the National AI Strategy becomes a catalyst or a cage for Ghana’s digital future.
The demo Aku AI is a government service chatbot, likely trained to answer questions in Local languages. Some say it was rather a video recording, because it couldn’t respond to the Honourable minister’s questions in real time, yet it was presented as an innovation.
But in Kumasi, a farmer already asks Farmerline’s voice assistant for weather updates in Twi. In Accra, a nurse uses mPharma’s AI to predict drug stockouts. At KNUST, students are running transformer models for liver ultrasound segmentation.
This backwardness is not technical, it is institutional. While Rwanda launched its National AI Policy in 2023 with a $30 million fund and CMU-Africa as anchor, Senegal embedded AI in its “New Deal Technologique” in 2022, Ghana waited until April 2026 to launch a strategy document.
REALLY?
Mandatory AI adoption for government agencies does not even start immediately, but later in 2026. Are we serious? In fact, is our government serious? Listen, in the language of technology, that waiting period is a generation.
Aku AI is therefore not a breakthrough. It is a clear admission of our backwardness. It admits that Ghana’s public sector has been outpaced by its own citizens. The same Minister for Communication, Digital Technology and Innovations Honourable Samuel Nartey George while addressing a stakeholder session on AI readiness at the Ministry of Finance, Aug 2025 said that recent breaches had targeted multiple state portals, including attempts to alter human resources databases. He linked these incidents to lax cyber practices, such as logging into official email accounts on unsecured public devices. Aku AI cannot mask that reality. The state is not leading Ghana’s AI revolution. It is chasing it.
Backwardness is not destiny. It can be leveraged, if Ghana uses the National AI Strategy to do what only government can do.
“Only 17.5% of BECE candidates passed all subjects in Gushegu.
This is not just a statistic—it’s a system failure.
Read the full report: https://simbaghana.org/wp-content/uploads/2025/10/IAA-SAGS-SCHOOL-MONITORING-REPORT_GUSHEGU-2.pdf
First, data: The private sector has models. The state has the data. For instance, the NHIS holds the health records of 17 million Ghanaians; ECG holds real-time energy consumption; GRA holds tax patterns. None of it is accessible to the many Ghanaian startups building AI today. The strategy’s commitment to standardise data through the National Data Centre and the piloting of AI in five ministries is the first real chance to unlock public data for public good. Aku AI is primitive, but if it forces the Finance Ministry to digitize workflows, then the next version of Aku, built by a Ghanaian startup, not a ministry IT department, can be transformative.
Second, procurement: The strategy makes AI adoption mandatory for over 100 MDAs from 2026. That is the single largest domestic market creation event in Ghana’s tech history. If procurement processes should transparent and local, so that the startups currently serving telcos and local farmers will finally have a government customer. The One Million Coders Programme launched by the Minister and the Girls-in-ICT Programmes can produce the talents. And the Al strategy can provide the demand.
Third, sovereignty: Local language AI is not a demo feature. It is an economic infrastructure. NLP Ghana, AkanNLP, and Khaya AI have already proven Twi, Ewe, and Dagbani speech systems work. A national strategy can make them the default for citizen services, ensuring Ghana’s AI future is not mediated by English or foreign models. That is a possibility no startup can guarantee alone. Only policy can.
Possibility is not guaranteed. The same strategy that can unleash Ghana’s AI industry can also strangle it. Three risks stand out.
1. Regulatory overreach: The Emerging Technologies Bill is being drafted, as the Minister posited, to “establish standards” for AI, robotics, and blockchain. Standards are necessary. Licensing regimes are not. If the Bill demands government approval for every model deployment, Ghana will do to AI what it once did to fintech: stall it, until Kenya and Nigeria own the market.
2. State competition: Aku AI is harmless as a demo. Aku AI v5, built by for example NITA with a 200-person team and subsidized compute, is a threat. The state must not become a vendor. Its role is to be a customer and a referee, not a player. Every cedi spent building in-house chatbots is a cedi not spent buying from Accra’s AI startups.
3. Infrastructure nationalism: The Bank of Ghana is already encouraging localization of financial workloads. The AI Infrastructure Readiness session stressed data sovereignty and trusted cloud environments. If this hardens into a rule that all AI inference must run in Ghana, startups will face costs that make them uncompetitive regionally. Equinix Ghana and Digital Realty will benefit. A two-person NLP team in Kumasi will not. Sovereignty without pragmatism is self-sabotage.
The final risk is cultural: mistaking symbolism for substance. Aku AI is a symbol. So was Hope City. Ghana does not need another ribbon-cutting. It needs procurement contracts, open datasets, and a Bill that protects citizens without punishing builders.
Ghana’s National AI Strategy is necessary precisely because, the demo Aku AI is backward. The strategy is late, but lateness is not fatal if it is followed by speed. The AU estimates AI could add $1.3 trillion to Africa’s GDP by 2030. Ghana will not claim its share by demoing yesterday’s chatbot. It will claim it by doing three things before 2027:
1. Open the data: Government entities such as NHIS, ECG, GRA, and Births & Deaths must release anonymised APIs for licensed Ghanaian companies.
2. Buy local, first: The 2026 mandatory adoption envisaged by the National AI strategy must include a Ghanaian-built preference clause for all MMDAs.
3. Regulate lightly: The Emerging Technologies Bill must target harm; bias, privacy breaches, deepfake fraud, not innovation itself.
The Demo Aku AI today is a symbol of how far the government has to go. It can still become the symbol of how fast Ghana can move. The difference between backwardness and leadership is not the model you demonstrate. It is the market you create.
The strategy has been launched. The bill is being written, I believe. The risk now is not that Ghana has no AI policy. The risk is that it may have the wrong one.
Abdul-Mumin Sofo Yumzaa, Executive Director, Simba Ghana
Email: ar.mumin@simbagha.org
About the Author
