Training the state in AI: the AfDB, Google and Cameroon launch programmes for African civil servants

For years, the digital transformation of African administrations was measured mainly in equipment: data centres, networks, software, online platforms. September 2026 marks a shift. Within a few days, two initiatives put the emphasis on something else: the skills of the women and men who make the state work.

In New York, on the sidelines of the 81st United Nations General Assembly, the African Development Bank (AfDB) launched, with UNDP, Google and the Apolitical platform, a free artificial intelligence training programme for public servants across the continent. In Yaoundé, on 21 September, the Cameroonian government opened a national platform to train 15,000 civil servants in digital skills, from AI to electronic archiving.

The continental programme: courses from Stanford and Bocconi

The AfDB programme is part of Apolitical’s Government AI Campus initiative, a learning platform for civil servants. It offers three courses: AI fundamentals for public servants, with its opportunities, risks and possible strategies; AI leadership in government, to build trust, skills and a culture of innovation; and smart public procurement with AI. The content was designed by Stanford Online and the SDA Bocconi School of Management in Italy.

The training is self-paced and free. It is open to serving officials in the AfDB’s regional member countries who work on AI in public services, digital infrastructure, public-private partnerships or innovation. Google and Apolitical provide the platform and part of the content. The UN partner, the AI Hub for Sustainable Development, is a G7-backed initiative co-led by UNDP and the Italian Ministry of Enterprises, and linked to Italy’s Mattei Plan for Africa.

« AI is changing the way we work. It promises major productivity gains if its users fully understand how to use it. Building that understanding within governments has been neglected, and this partnership helps us put that right. »

Ousmane Fall, director of industrial and trade development, African Development Bank (paraphrased)

The AfDB presents this training as the first implementation of the recommendations in its report on AI productivity gains in Africa, which estimated that an inclusive rollout of AI could add up to $1 trillion to the continent’s GDP by 2035. The report identified skills and trustworthy governance among the five conditions needed to get there.

The Cameroonian programme: 50 days, five modules

In Cameroon, the train.patnucelearning.cm platform is part of the digital transformation acceleration project (PATNUC), overseen by the Ministry of Posts and Telecommunications and funded by a 42 billion CFA franc World Bank loan. The training takes place entirely online over 50 days, with ten days per module, so that staff can train without leaving their posts.

• Artificial intelligence and blockchain: understanding these technologies and their uses in government.
• Digital transformation and service to users: simplifying procedures for citizens and businesses.
• Digital project management and agile methods.
• Change management.
• Archiving and electronic document management, to support the move to paperless processes.

The programme is open to staff from every ministry, including those in regional offices, as well as employees of public institutions and state-owned companies. Participants who complete the course receive a certificate.

« Technology alone does not transform the administration. The administration rests above all on the women and men who run it, whose skills must evolve at the same pace as digital tools, methods and practices. »

Minette Libom Li Likeng, Cameroon’s Minister of Posts and Telecommunications (our translation)

Platforms already built, still little used

PATNUC’s national coordinator, Winfred Mfuh Fuaye Kenji, sums up the challenge: having infrastructure and platforms is not enough; their users need to know how to use them. The project already claims more than fifteen digital platforms and infrastructures, including an electronic voucher subsidy system for the agro-pastoral sector, an agricultural statistics platform, the digitisation of procedures in two ministries, and two tools developed with the telecoms regulator to monitor the frequency spectrum and measure the quality of mobile networks.

This is a frequent admission in donor-funded digital transformation projects: the tools are delivered, but adoption remains low, for lack of training, support and sometimes the will to change habits. A civil servant who has not mastered a platform will keep working on paper, and the investment will remain underused.

Initiatives are multiplying

These two programmes are not isolated. In several countries, digital ministries are launching their own online academies, often with the support of big tech companies, which also see them as a way of familiarising administrations with their tools. Kenya, Rwanda, Nigeria and Morocco have announced digital and AI skills training for their officials or young people in recent months. This proliferation has a downside: the risk of duplication, uneven content and certificates without common recognition from one country to another.

The presence of tech giants in these programmes also raises a question of independence. Training designed with a vendor naturally tends to showcase its tools. Administrations have an interest in demanding commercially neutral content that teaches people to evaluate and compare solutions rather than adopt one out of habit. Using universities such as Stanford and Bocconi to design the AfDB’s courses is a step in that direction.

What will make the difference

Online training has an obvious advantage: it reaches many people at low cost. It also has a well-known weakness: completion rates. Many enrolees drop out along the way, especially when the training comes on top of a normal workload and the connection is expensive or unstable. Success will therefore be measured by the number of officials who complete the course, and above all by what they do with it afterwards.

For these programmes to change how the state works, three conditions seem necessary. The training must lead to concrete uses in daily work, such as tools to help with drafting, translation or case processing. Trained officials must be recognised, in their appraisals or their careers. And there must be clear rules on what a public servant may entrust to an AI, particularly regarding personal data and official secrecy.

Artificial intelligence is arriving in African administrations whether they are ready or not: through the tools officials already use on their phones, through vendors’ offers, through agreements signed with the sector’s big companies. Training civil servants in its uses and risks is not a luxury. It is the condition for the state to remain in charge of its decisions, rather than unknowingly delegating them to tools it does not understand.