East Africa Just Approved a Shared AI Module for Every Degree. Adoption Is Still Voluntary.

East Africa now has a regional answer to a question every university on the continent is being asked: what should every graduate know about AI, regardless of what they studied?

On October 31, at the ministerial session of the 21st Sectoral Council on Education, Science and Technology, Culture and Sports in Kampala, Uganda, EAC ministers adopted three strategies at once: the EAC AI Strategy, the EAC Digital Transformation Strategy, and the EAC Biotechnology Strategy. The meeting ran from October 27 to 31 at the newly completed headquarters of the Inter-University Council for East Africa (IUCEA). The strategies were developed by the EAC Secretariat with the East African Science and Technology Commission (EASTECO), and they cover education, agriculture, healthcare, industry, and public administration.

The item that matters most for practitioners is the smallest one in the announcement, and the most concrete. Ministers approved an AI Foundational Module for Higher Education Institutions, intended to equip undergraduate students with essential knowledge and practical competencies in AI, irrespective of their academic discipline. Partner states were directed to consider its adoption across university degree programmes, while IUCEA was tasked with developing a regional implementation framework to guide the rollout.

What was actually approved

The October 31 package has five parts worth separating, because they are not all the same kind of commitment:

1. Three strategies. The AI, digital transformation, and biotechnology strategies give the region a common framework. A strategy is a direction, not a delivery date.
2. One shared module. The AI Foundational Module is a curriculum object: the same foundational AI teaching, designed to work across degree programmes and disciplines. That is the piece a lecturer can actually teach from, once the framework exists.
3. One network, made operational. Ministers approved the establishment and operationalisation of the Regional AI Network of Universities, to strengthen collaboration in AI education, research, and innovation. This network is not new in concept: the EAC AI Alliance launched it in Kigali in March 2026, and it formally launched in Nairobi on July 29, 2026, with universities invited to serve as national nodes. The October decision moves it from launch to operation.
4. Three teacher instruments. The council endorsed three regional AI instruments for teacher education, covering curriculum and competency, institutional policy, and a code of conduct. This is the part most announcements skip, and it is the part that decides whether a module survives contact with a real classroom. A shared module without trained teachers is a PDF.
5. Kiswahili benchmarks. Ministers directed the development of AI certification benchmarks for Kiswahili, and adopted the EAC Kiswahili Language Policy for submission to the Council of Ministers. Kiswahili is the region's shared language, and a benchmark for it is a testable claim: a model either performs in Kiswahili at a stated level, or it does not. That is a better foundation for language work than a press release claim.

Alongside the AI items, ministers approved a Technical and Vocational Education and Training (TVET) Qualifications Framework, to make qualifications comparable and portable across partner states. That matters for the same reason the module does: East Africa's bet is on people moving, studying, and working across borders, and portable credentials are part of that bet.

What was not approved, and why the wording matters

Read the verbs in the EAC's own release, because they carry the whole story:

• Strategies were "adopted". That is done.
• The module was "approved". That is done, as a module.
• Partner states were directed to "consider" its adoption. That is not adoption.
• IUCEA was "tasked" with developing a regional implementation framework. That framework does not exist yet in public.
• The council "acknowledged financial and resource constraints affecting programme implementation" and thanked development partners, including Germany, the European Union, and the World Bank, for supporting regional initiatives including the EAC AI Alliance.

There is no rollout deadline in the release. There is no new funding figure. There is no list of first universities, and no pilot cohort named. An earlier October review, reported by Okay News on November 1, described a course for technical and non-technical students covering practical AI use and how to assess AI-generated material critically, with a proposed pilot process that would select institutions by readiness and examine teaching approaches, lecturer preparation, and student participation before wider adoption. That pilot logic is sound. It is also still a proposal: the October 31 approval does not establish that classes or pilots have begun.

Why a shared module is a good idea, even with the caveats

For data and AI practitioners, the interesting part is not the strategy. It is the standardisation problem the module tries to solve.

First, a shared foundational module means a graduate in Kisumu, Kigali, or Kampala can be assumed to have met the same core material: what a model is, what training data does to an output, how to check an AI-generated answer, and where AI should not be trusted. Employers and postgraduate programmes currently cannot assume any of that. A common module, if actually taught, turns a guess into a credential.

Second, the teacher instruments matter more than the module text. The three instruments cover curriculum and competency, institutional policy, and a code of conduct. That is the right trio. Curriculum without policy leaves lecturers to improvise rules on AI use in assessment. Policy without a code of conduct leaves academic integrity to individual judgment. A region that ships all three together is thinking about delivery, not just documents.

Third, the Kiswahili benchmarks are the most testable part of the package. African language work keeps running into the same gap: a model is claimed to support a language, with no public benchmark to check the claim against. A regional certification benchmark for Kiswahili, if it is published with test data and scoring rules, gives builders, buyers, and ministries a number to argue about. If it stays internal, it is just another announcement.

Fourth, the network structure already exists. The Regional AI Network of Universities was designed as a distributed set of specialised nodes, not a single centre. That fits the region: universities differ widely in staffing, infrastructure, and digital capability, and a node model lets a strong institution anchor work that weaker ones can join. The October decision to operationalise the network is the bridge between the March launch and actual teaching.

Five things to watch

1. The IUCEA framework. When it is published, does it name a start date, a pilot cohort, and who pays for lecturer training? A framework without those three is a document, not a plan.
2. The first adopters. Which partner states move first, and do they adopt the module across degree programmes or confine it to computing degrees? The release says "irrespective of their academic disciplines". Watch whether practice matches that.
3. The teacher instruments in practice. Are the curriculum, policy, and code of conduct documents public, and do universities actually change assessment rules when they adopt them?
4. The Kiswahili benchmark spec. Is it published, with data and scoring, or announced and then left vague? A benchmark that nobody can run is not a benchmark.
5. The funding line. Does the next EAC budget, or a partner commitment, put money behind lecturer training and the network nodes? The October 31 release thanks existing partners. It does not announce new money.

In favour

A shared module is cheaper to build well once than eight national modules built badly. A regional network lets universities share lecturers, materials, and benchmarks instead of each solving the same staffing problem alone. And Kiswahili benchmarks, if published, would be a genuine public good for language technology in the region.

Against

"Consider adopting" is the weakest directive a council can give, and the release pairs it with an open acknowledgement of financial constraints. Universities with different staffing and infrastructure will not deliver the same module the same way, and a shared syllabus taught badly in half the region could do more harm than no syllabus at all, by certifying skills students do not have. The strategies also bundle AI with digital transformation and biotechnology, which spreads attention across three large agendas at once.

What would make me wrong

If IUCEA publishes a framework within months with a dated pilot cohort, named universities, and funded lecturer training, and if two or more partner states adopt the module across disciplines in the next academic cycle, then the cautious reading here was too cautious, and October 31 was a delivery decision, not just a direction. If, a year from now, the framework is still unpublished and no university can point to a taught cohort, then the announcement was the product, and this piece was too generous.

Related on Everyday Data Science

• Africa's AI Fortnight Was About Small Models and Speech Data. That Is the Point.
• Three Plans to Share Africa's AI Computing Power Landed in Six Days. The Hard Numbers Are Still Missing.
• Sovereign AI Got Hardware This Week: Two Deals, Two Very Different Tests

Sources

• EAC Ministers Adopt Regional AI and Digital Transformation Strategies, Approve Establishment of Regional AI Foundational Module for Universities (EAC press release, October 31, 2026)
• East African Ministers Approve University AI Module and Strategy (Okay News, November 1, 2026)
• EAC AI Alliance Invites Universities to Become National Hubs in Regional AI Network (EAC, March 31, 2026)
• Regional AI Network (AI4EAC, official launch July 29, 2026, Nairobi)

What would you teach first in a foundational AI module for students who will never become AI engineers, and what would you leave out?