Zanzibar’s AI Call for Public Servants Needs Contract Power, Not Just Ethics

In Zanzibar, First Vice-President Othman Masoud Othman opened the 11th APS-HRMnet conference in 2026, bringing together delegates from 21 African countries to discuss human resources and artificial intelligence in public service. The region’s leaders now need to turn their talk of ethics and youth into bargaining power over AI infrastructure and contracts.

Zanzibar wants AI-ready civil servants, not automated bureaucracy

Othman framed artificial intelligence as a support to human judgement rather than a replacement, stressing that institutions rely on competent, disciplined public servants committed to integrity and responsible service. He warned that digital reforms must be matched by good governance, transparency and accountability, and urged governments to set clear AI policies to protect citizens’ rights, secure information and ensure lawful use. He said that artificial intelligence should be a tool that helps public servants perform their duties more efficiently and must never replace their thinking, wisdom, professionalism or responsibility in decision-making, underscoring that human beings must retain authority over decisions.

The APS-HRMnet meeting, themed around strategic human-resource transformation through AI, brings together public-service managers who are being told their role now extends beyond recruitment and salary administration. Organisers say the agenda centres on reforms to public-service staffing and the use of AI in delivering services, in keeping with the event’s theme. Othman pressed participants to anticipate labour-market demands, identify new skills and understand AI’s implications for government employment. Minister Ridhwan Kikwete highlighted that around 1,200 participants would assess how public institutions can keep pace with technological change, and Suleiman Masoud Makame emphasised Zanzibar’s efforts to strengthen public service systems.

Africa’s youthful workforce is useless if AI authority sits abroad

For Othman, Tanzania’s youthful profile is an asset: national statistics show that those aged 15–35 make up 34.5 per cent of the population and around 40 per cent of public service staff. He called for inclusive digital programmes reaching women, persons with disabilities and rural public servants. That ambition is sound, but a young, digitally aware workforce is only an advantage if those officials can interrogate and control the AI systems embedded in their ministries.

Without institutional authority over data, models and service continuity, African governments risk training a generation of officials who operate foreign-run systems they cannot properly evaluate. A United Nations report released in September underscored how weak Africa’s starting position is: as of June 2025, high-income economies held 77 per cent of global co-location data-centre capacity, while low-income countries had less than 0.1 per cent. Against that backdrop, a youthful civil service on its own does not deliver sovereignty or bargaining power. It must be tied to concrete rights to test performance, safeguard data and rehearse exits.

World Bank advice shows contracts, not workshops, decide who controls public AI

Across the continent, AI-linked partnerships are taking shape. A September letter of intent between South Korea and the African Development Bank set out plans for a joint AI hub, while the U.S. Trade and Development Agency backed a feasibility study for AI-ready data centres in Nigeria. If carried through, each of the two projects is expected to draw in investment, specialist expertise and much-needed infrastructure. Yet the descriptions offer no indication of whether host institutions would be able to reopen the agreements once their economies or public services have begun to depend on those systems.

The World Bank unveiled its inaugural digital strategy on 20 October. In its World Development Report 2026, the bank advises lower-income economies to begin with the tools they already have, adapt them to local conditions, and progress to more sophisticated technologies only when their resources can justify the expense. The sequencing makes sense for development, but every step raises a governance question: which party holds the data, documentation and operational knowledge needed for the next upgrade? Those disparities mark an unequal starting position rather than proof that foreign partnerships are exploitative; the real risk comes when governments gain access to AI services without the means to test providers’ claims, defend legitimate interests in local data, maintain essential operations or switch suppliers on clearly understood terms.

Othman is right to insist that AI deployment in the public sector must be guided by accountability and ethical standards. Yet ethics workshops alone will not give Tanzanian or Zanzibari ministries the authority to demand documentation, reproduce performance claims in the languages, populations and operating conditions they serve, and insist on secure access to their own records. APS-HRMnet’s members are being told to become strategic partners in institutional development; that role should now extend explicitly to procurement language, data governance and continuity planning for every major AI-enabled system.

African Development Bank alliances must embed retained capability, not just share enthusiasm

Some will argue that African governments first need basic connectivity and functioning services, and that loading contracts with conditions risks higher prices, fewer suppliers and slower delivery. In countries with limited electricity or technical expertise, buying mature cloud services or foreign-developed models may indeed yield more benefit than attempting to recreate a frontier laboratory. The Eurasia Review piece itself notes that technology exchange can be mutually beneficial.

But dependency is only tolerable when it is priced and tested. Egypt’s Cabinet Information and Decision Support Center warned on 5 November of the financing risk around African data-centre projects, including who pays for electricity and supporting infrastructure if demand diverges from forecasts. That is precisely the kind of retained-capability discipline Africa’s public-service managers need to internalise. Regional bodies already in motion, such as the African Development Bank’s partnership with South Korea and the African Union’s Agenda 2063 framework, should embed four simple tests into public-sector AI deals: local rights to evaluate systems, lawful control over data and decision records, rehearsed continuity and exit plans, and institutional learning that leaves ministries with working evaluative capacity, not just certificates from short courses.

The United Nations’ Global Dialogue on AI Governance, which brought 163 delegations to Geneva in July 2026 and will reconvene in New York next May, offers a forum for common principles. APS-HRMnet, representing human-resource leadership from 21 African countries in Zanzibar, is better placed than most to turn those principles into day-to-day practice. It should draft practical retained-capability schedules for public-sector AI deployments and press the World Bank and African Development Bank to tie future funding to governments that can audit consequential results, retrieve their data and compare alternative suppliers in reality, not just on paper.

Why it matters

Zanzibar’s initiative is likely to favour African public servants and APS-HRMnet by giving HR managers support for claiming a more strategic role and stronger emphasis on skills development. It also aligns with the African Development Bank’s AI hub plans and the World Bank’s digital strategy, and with United Nations and African Union efforts tied to Agenda 2063 and global AI governance. The immediate financial burden falls on African treasuries funding training and digital systems, and on residents whose data will underpin public-sector AI, with infrastructure and electricity costs looming over data-centre initiatives such as the Nigeria feasibility study and the risks highlighted by Egypt’s Cabinet Information and Decision Support Center.

For developing countries the main risks are dependency on foreign clouds, weak leverage at contract renewal, and civil services trained to operate systems they cannot fully audit or move. The tangible commitments in play are those with dates and named programmes: the African Development Bank–South Korea letter of intent on an AI hub, the U.S. Trade and Development Agency study in Nigeria, the World Bank’s 20 October digital strategy, the UN’s quantified data-centre imbalance and the APS-HRMnet gathering in Zanzibar. Promised benefits beyond those documents rest mostly on institutional assurances. APS-HRMnet should now work with the African Development Bank and World Bank to publish a shared model for retained capabilities in public-sector AI contracts, so parliaments and citizens can later measure whether governments kept real control over data, evaluation and continuity.