| Getting your Trinity Audio player ready... |
Following his participation in the 2026 Alliance for African Partnership Annual Meeting in Dar es Salaam, Tanzania, ANAPRI Director of Programmes and Policy, Dr Tinashe Kapuya reflects on what the AI transition means for Africa’s workforce, institutions and agricultural policy research.
One of the most revealing moments at the public dialogue on AI and the future of work concerned a powerful computer that a university had acquired but was not ready to use.
Professor Idrissa Sarr of Université Cheikh Anta Diop described an institution that spent months procuring advanced equipment, only to find that its specialists and IT team could not make it widely accessible. Students developed promising applications but could not fully explain the results, while a hospital lacked the infrastructure to deploy one application. More tellingly, the applications required only a fraction of the computing power the university had purchased.
His lesson was disarmingly simple: “Make sure you are buying what you can use.”
That warning captured a central message of the Alliance for African Partnership–Andrew Carnegie Foundation dialogue, held at the University of Dar es Salaam on 17 September: Africa’s challenge is not simply to acquire AI, but to build institutions capable of choosing, governing and applying it to real problems.
Bruce Mwile Mwasenga, Chief Operating Officer of CRDB Bank Plc, identified five capabilities for tomorrow’s workforce: digital and data literacy; AI fluency; critical thinking and problem-solving; ethical leadership; and lifelong learning. These are becoming essential capabilities not only for computer scientists, but also for economists, bankers, agronomists, lecturers and public officials.
The banking example made the shift tangible. CRDB has identified 36 AI use cases, including an application that can review a 300-page contract in minutes rather than days. Yet while AI may accelerate the reading, people must still interpret risk, exercise judgement and remain accountable. The future of work will be shaped not only by whether occupations disappear, but by how tasks within them are reorganised.
Professor Laura Lee McIntyre of Michigan State University argued that universities must prepare students to work with AI without outsourcing learning itself. Employers still need people who can communicate, collaborate and adapt. Critical thought becomes even more important when a machine can generate a plausible answer in seconds. AI literacy, therefore, must spread beyond engineering and computer science.
Makerere University Vice-Chancellor Professor Barnabas Nawangwe placed the debate within Africa’s employment challenge. Universities must become engines of research, innovation and entrepreneurship, while Africa must adopt and adapt AI to solve African problems rather than simply reproduce the technology frontier on someone else’s terms.
For ANAPRI’s centres, the discussion points to a simple starting principle: begin with problems, not products.
First, centres could identify a small portfolio of high-value, low-risk applications; from cleaning agricultural datasets and improving literature reviews to comparing policy scenarios, translating research products and making technical findings more accessible. These should be tested through supervised pilots, with sensitive data, causal interpretation and final policy recommendations remaining subject to human oversight.
Second, capacity-building should reflect different needs. All staff require foundations in data literacy, verification, privacy and responsible use; researchers may need deeper technical skills; and managers need capabilities in strategy, procurement, cybersecurity, intellectual property and data governance. A generic one-day workshop will not achieve this. Learning should be continuous and tied to real work.
Third, ANAPRI could draw on its network advantage through shared training, common data standards, reusable tools and code, communities of practice and, where appropriate, pooled access to computing infrastructure rather than standalone investments.
Fourth, responsible use must be designed in from the start. Clear rules are needed on confidential data, disclosure of AI assistance, source verification, bias, authorship and human sign-off. Success should be measured not by licences or training sessions, but by whether AI saves time, improves analytical quality, widens access to evidence and strengthens policy influence.
The lesson from Dar es Salaam is not that African institutions need more AI at any cost. They need the capacity to determine where it adds value, the systems to use it responsibly, and the people capable of questioning its outputs. Across Africa, this will require collaboration among universities, research institutions, governments, the private sector and diverse networks.
For ANAPRI, as a member of the Alliance for African Partnership, this conversation continues. Ahead of the 13th ANAPRI Annual Stakeholders Conference, a pre-conference workshop on 26 October 2026 will focus on The Political Economy of AI in African Agrifood Systems. The conversation will continue during the main conference, taking place from 27–29 October in Addis Ababa, Ethiopia, with a dedicated session examining AI and food systems in Africa. Ultimately, the opportunity is not simply to adopt new technologies, but to build the people, institutions and partnerships needed to apply them responsibly to the problems Africa needs to solve.