Ghana's national AI strategy: A plan to expand AI education, digital infrastructure, and data access

The problem

By the early 2020s, Ghana had a growing digital and start-up ecosystem, expanding mobile connectivity, university programmes in computing and data science, and international AI investment, including Google's AI centre in Accra and local start-ups working on AI for health and agriculture.

That activity was fragmented and uncoordinated. Universities, companies and public-sector teams worked on AI in isolation, with no shared platform to connect them or pool what they learned. Awareness of what AI could do was low across government, and there was limited political will to prioritise it. The country had few local datasets that were accurate, representative, and accessible across sectors, and the data that did exist remained within the institution that collected it. Digital infrastructure was uneven, with an average 4G coverage reaching 41% in rural areas against 88% in urban areas. Skilled AI practitioners were in short supply. AI models built on data from other countries often produced inaccurate or biased results when applied to Ghanaian conditions. It became increasingly important to connect these efforts, align the actors involved, or set a shared direction for the responsible adoption of AI.

The solution

The strategy is organised around eight pillars, grouped into four ecosystem enablers and four accelerators. The enablers cover expanding AI education and training, empowering youth for AI jobs, deepening digital infrastructure and inclusion, and facilitating data access and governance. The accelerators cover coordinating the AI ecosystem, increasing adoption across sectors such as healthcare, agriculture, financial services and transportation, investing in applied research, and promoting AI adoption in the public sector.

Several commitments are specific. The AI Ready Ghana programme aims to train more than one million young people by 2033 and at least 10,000 AI researchers. An NLP Centre of Excellence (NLP refers to natural language processing, the field of AI that works with human language) is planned to support AI in Ghanaian languages. The strategy calls for GhanaChat, a large language model (an AI system trained on large amounts of text) built on government data for internal public-sector use, intended to support staff productivity while keeping confidential information off foreign platforms. It also sets out plans for a National Deep Science Institute, a Ghanaian artificial general intelligence lab and ten AI companies valued at more than a billion dollars by 2035.

To drive delivery, the strategy recommends establishing an independent Responsible AI Authority within the first year, supported by a National AI Office to coordinate AI policy and implementation across ministries. At the launch event, the approach was demonstrated through Aku, an AI assistant that fielded questions in six languages, English, Ga, Twi, Dagbani, Ewe and Gonja, and ran through scenarios that included helping a rural clinic with diagnosis, a financial-services task, and everyday support for farmers and fishermen.

The impact

The strategy is Ghana's first comprehensive national AI framework, dated December 2025 and launched in April 2026. Its development engaged around 40 experts and four public-sector workshops, and produced a SWOT analysis of the AI ecosystem and an action plan with named responsible actors, timelines and targets across all eight pillars.

As a framework rather than a deployment, outcome data is not yet available, and the figures below are targets that the strategy sets, not results.

AI is targeted to contribute GH¢200 billion to GDP by 2030 and GH¢500 billion (around US$45 billion) by 2035. The strategy also aims to attract GH¢200 billion in investment by 2035, make one million young people AI-ready by 2033, and curate one trillion tokens from Ghanaian datasets by 2030. Ghana is working with UNESCO to apply the AI Readiness Assessment Methodology, which is intended to provide an evidence base to guide implementation.

Lessons learned

• Grounding the strategy in wide consultation kept it tied to the ecosystem's own strengths and gaps. The plan was built from around 40 expert interviews and four workshops spanning government, academia, start-ups, the private sector, and civil society, and the SWOT analysis drawn from that engagement serves as the stated basis for its recommendations. A government planning a similar framework can use a consultative process to anchor it in the country's actual standing.

• A SWOT analysis gave the strategy a clear set of weaknesses to address alongside its strengths. The diagnostic named constraints, including low political will, uneven rural infrastructure, gaps in data governance, and the risk that AI models trained on foreign data may not fit the local context, became the problems the recommendations were written to address.

• Building ethics and oversight into the strategy keeps them part of implementation rather than a separate step. The strategy aligns with the UNESCO ethics recommendation and establishes the need for a Responsible AI Authority to ensure responsible deployment through delivery. At the launch, the Speaker of Parliament said the country must be morally grounded as well as technologically advanced, and the Chief Justice said the technology must remain subject to the rule of law.

• Treating data sovereignty as a design choice allows a government to open up to investment while retaining control over sensitive data. The strategy pairs openness to international investment with the development of local datasets in Ghanaian languages, the domestic hosting of data, and the use of a government large language model for sensitive work.

Launch year: 2025

This case study was written with assistance from artificial intelligence.