The ThursdayBriefing 8 October 2026
By Dr Blessing Ivan VavaDonald Trump’s administration was clear about whose future it was celebrating. As the US president gathered technology executives at the White House on September 29, his officials hailed a coming “Golden Age”. The same day, Trump signed an executive order directing federal agencies to use “super intelligence” instead of “artificial intelligence” in official communications, to the extent permitted by law. Watching from Africa, I was struck by how plainly the ambition was stated. These companies may serve the world, but the political project has a national address.
The gathering produced the White House Accord on Super Intelligence, signed by Trump and leaders from Google, Anthropic, Meta, OpenAI, xAI and Nvidia. It sets out voluntary commitments to internal controls, external audits and board oversight. Trump called it “morally binding”. The document leaves open the possibility of legislation later.
Decisions made by the US government and major AI suppliers will shape which systems African health and education ministries, public hospitals and schools can buy, and on what terms. The accord names cybersecurity, biosecurity and chemical threats among its concerns. It offers no explicit route to redress for a patient harmed by an AI-assisted decision or a child poorly served by an AI tutor. African governments should ensure that national laws and procurement contracts protect those using these systems. Health and education authorities should enforce standards, while independent regulators and courts provide avenues for complaints and remedies.
Trump knows what he wants AI to deliver for America. The harder question is whether African governments know what they want before they sign a technology agreement.
Washington has put its intentions in writing. A July 2025 executive order seeks worldwide adoption of American AI technologies, standards and governance models. It backs exports of complete technology packages, from computing infrastructure to applications, and explicitly aims to reduce dependence on technologies developed by US adversaries. Selling technology is part of a strategy for extending American power. At the White House meeting, Trump put the rivalry plainly: “We have a very big lead, and we’re going to keep our lead.” He has also resisted sweeping federal regulation on the grounds that it could slow American innovation and benefit China.
China’s Global AI Governance Action Plan offers another route, promising open-source cooperation and support for the Global South. Those promises deserve the same scrutiny as Washington’s. Open-source tools may be difficult to put to lasting use if local institutions lack the expertise to adapt them or the money to maintain them. Still, credible alternatives from Chinese suppliers could give African governments more room to negotiate better terms with American companies.
That leverage is greatest before a contract is signed and can vanish soon after. Suppose Zimbabwe's Ministry of Health and Child Care buys an AI system to support patient care in public hospitals. Once patient records and daily services run on that supplier's software, replacing it requires costly data transfers and retraining staff. If fees rise, the ministry may simply pay, because switching would disrupt treatment. One procurement decision can narrow the choices of future governments.
Such dependence can be built into contracts, where a government’s need to improve public services meets a supplier’s commercial interests. I examine these encounters through geosociotechnopolitics, a framework that connects local needs and political choices with the ambitions of powers such as the US and China. African officials may have limited bargaining power and face pressure to secure investment quickly. They still need to establish who controls the system and where the data goes, and ensure that citizens have someone to hold to account.
These questions surfaced in Zimbabwe’s 2018 facial-recognition partnership with CloudWalk. Contemporary reporting described plans for Zimbabwean facial data to help train the Chinese company’s systems. MISA Zimbabwe raised concerns about the anticipated transfer and called for parliamentary scrutiny. Those reports do not establish what data was ultimately transferred. They do show why consent and oversight belong in the discussion before a partnership proceeds.
The supplier’s nationality cannot settle these questions. An American product carries no automatic guarantee of accountability, just as a Chinese agreement does not by itself establish dependence. The terms of the deal matter, along with the institutions responsible for enforcing them.
Connecting schools and training teachers to use an AI tutor may leave an education ministry dependent on software it cannot properly assess or adapt. Teachers might notice that it gives poor answers in Shona, but who can investigate the problem and get it corrected? That requires local expertise and an agreement that gives the ministry a say over how the system is used. Teachers should have time to test the tutor against the curriculum before purchase, and their findings should influence the decision. Otherwise, once schools rely on it, officials could struggle to resist higher fees or proposed changes to the use of pupils’ data. This is the governance divide that concerns me: expanding access without giving public institutions the means to protect the people using the technology.
For a district hospital in rural Malawi, an AI tool that helps interpret scans could ease pressure on scarce specialist services. Before accepting such an offer, hospital managers and clinicians would need evidence that it works reliably for their patients. They would also need to establish where patient records would be stored, who could access them and who would remain responsible for clinical decisions. If the supplier withdrew the service, the hospital would still need a way to assess and treat patients.
Africa has policy foundations for this work. The African Union’s Continental AI Strategy, endorsed in July 2024, calls for an Africa-centred approach and cooperation among member states. The Malabo Convention on cybersecurity and personal data protection entered into force in 2023. The AU also has a Data Policy Framework. Turning these commitments into effective oversight requires public institutions with staff, budgets and the authority to act.
Procurement offers an immediate starting point. Before signing a major agreement, a government should require evidence that the system works in its intended setting, clear limits on the use of sensitive information and a practical way to change providers. Independent scrutiny at this stage could prevent years of avoidable dependence.
Smaller countries need not shoulder every cost alone. The Southern African Development Community (SADC), the East African Community (EAC) and the Economic Community of West African States (ECOWAS) could help member governments share the cost of technical teams to evaluate AI systems and advise on contracts. Universities within these regions could pool computing resources and contribute researchers to those teams. Officials would then have access to expertise independent of the company seeking their signature.
Hospital staff might leave a training workshop knowing how to use an AI system, yet still have nobody locally to turn to when it produces questionable results. That is a limited return on the promise of capacity building. Paying domestic firms to develop and maintain applications would help build expertise that institutions could draw on long after the training ends. Universities also need sustained access to computing resources to test tools in Shona, Swahili or Chichewa. Much of that investment could be lost if public institutions cannot afford to keep trained staff. Governments should ask what their institutions will be able to do for themselves once the supplier’s team has gone home. The answer need not be a home-grown ChatGPT. It should at least include enough expertise to recognise a poor offer and negotiate a better one.
Citizens also need protection from their own governments. Putting an AI system under national control does not, by itself, prevent officials from using facial recognition to track critics or algorithms to deny public services without explanation. People need a way to challenge those decisions, backed by independent oversight and enforceable limits on state power. Otherwise, sovereignty may merely change who holds the power to misuse the technology.
Trump has put American power behind the AI industry. African leaders should be able to explain what their own agreements are for, and remain answerable for them years later. When a hospital’s system fails, it matters whether help is available locally. A university should have more to show for a partnership than staff trained to use somebody else’s software. Citizens have a right to know what these deals cost and who has access to their data, long after the officials who signed them have left office.
America is entitled to pursue its golden age. Africa’s task is to negotiate a future its own citizens have reason to welcome.
Dr Blessing Ivan Vava is a Zimbabwean researcher and Executive Director of the Southern Africa Coalition for Democracy and Accountability (SACDA). His work examines technology, sovereignty and democratic accountability in Africa. He holds a doctorate in Communication Studies from the University of Johannesburg, where he researched Chinese ICT investment in Zimbabwe, and writes The Thursday Briefing on his blog, The Vuvuzela Times.
