The crossroads may be the best road

In a seminar room at Future Africa, on the University of Pretoria campus, a senior South African scholar (climate scientist, IPCC veteran, decades in international rooms) paused mid-panel. She told us she used to be petrified to speak in places like this. Always petrified. And then:

Honestly, we don’t have time. And quite frankly, I don’t really care anymore. We have to move.

That mix of urgency and exhaustion captures something I have been hearing across four gatherings this past month. These include roundtables I attended at Future Africa in Pretoria, at the Center for Sustainable Development at Addis Ababa University, and at the University of Malawi in Blantyre; and the conversations that followed a public lecture at the University of Mauritius, where the explicit focus was artificial intelligence and global development.

Each had its own preoccupations. Energy sovereignty in Pretoria. The unraveling of the post-1945 order in Addis. The future of aid in Blantyre. The asymmetric position of small island states in Mauritius. But underneath every discussion, the same questions kept surfacing in different accents and disciplines. Who defines development? Whose knowledge counts? And what does African agency actually look like, in practice, in 2026?

The conversations were not tidy, and participants disagreed productively with each other and with me. These are threads worth paying close attention to.

The crossroads moment

The most insistent topic in Addis Ababa was the geopolitical moment. The dramatic cuts to American development aid. The shock to the humanitarian system. The retreat of multilateralism. European budgets redirected toward defense and Ukraine.

What I found particularly striking was how rarely this was framed as catastrophe. One Ethiopian participant called the disruption “a blessing in disguise.” Another said his country had been living “in warm water,” comfortable enough not to notice the temperature, and was now being forced to confront reality. A third put it bluntly: “This is a wake-up call.”

The same instinct echoed in Blantyre, where the stakes are sharper. Nearly forty per cent of Malawi’s national budget has historically come from development partners. The country owes the IMF something on the order of several hundred million dollars. There is no plausible way that debt gets repaid from current revenue.

So when participants in Blantyre talked about the future of aid, the talk had teeth. One NGO director put it this way:

The future is bright. But to when and by whom, that is the problem. We had Vision 2020. Now we have Vision 2063.

He was almost laughing as he said it. “Vision 2063 is a very convenient deadline, because most of the people making the promises will not be alive to be held to them.”

Another participant offered the line of the morning: “If we don’t do away with aid, then aid will do away with us.” Meaning: if we don’t take the lead in shaping what comes next, the collapse of the old model will simply consume us. As it happened, the front page of The Nation in Malawi that morning carried the headline “Malawi, EU shift focus from aid to production.” Whether that turns out to mean anything substantive remains to be seen, but the conversation is moving.

Not everyone was sanguine. In Addis Ababa, one scholar pushed back hard. “I’m not sure we can describe the collapse of the international order as an opportunity. How exactly can we benefit from a world in which there is no order, and might makes right?”

That tension (opportunity versus disorientation) ran through every gathering. The most useful frame came from a participant in Addis who reached for Gramsci. We are living through an interregnum. The old order no longer functions, but the new has not yet emerged. The post-1945 system, he argued, was never rule-based. It was power-based. The rules reflected the preferences of the victors of WWII. That hegemony is now genuinely cracking. And nothing has replaced it.

A younger Ethiopian participant gave us the line I have not been able to forget.

The crossroads may be the best road, precisely because it opens up multiple directions.

Everything that follows is, in some way, about what we choose to do at that crossroads.

Whose knowledge counts?

The question of African agency runs deeper than aid budgets. It runs into how knowledge is produced, who gets cited, and who feels entitled to speak.

The opening panelist in Pretoria (that climate scientist I mentioned above) made a striking observation. African scholars, she said, are not actually marginalized in the international conversation anymore. They are in the room. The problem is more troubling. They are in the room, and they don’t speak, because they aren’t sure they have anything powerful to say.

She quoted the geographer Farhana Sultana, who asks a simple question: “Have you cited a black woman today?” The answer, Sultana argues, changes the discourse. This was not a generic decolonization speech. It was specific. The panelist named buzzwords she is trying to retire from her own vocabulary. End users, because, as she put it, “we are not end users; we are all users.” Resilience, because the language too often turns structural neglect into a virtue.

She referenced Maria Kaika, the urban political ecologist at the University of Amsterdam, whose paper “Don’t Call Me Resilient Again” makes the case directly. After Hurricane Katrina, a New Orleans resident put it bluntly: “Do not call me resilient again. We are conditioned to be in this place where we need so-called resilience only.” Kaika’s argument, and our panelist’s, is that the language of resilience papers over the failure to address the causes that created the vulnerability in the first place. And the rush to be transdisciplinary, to bring everyone into the room, can sometimes become a way of looking inclusive while flattening the asymmetries that actually matter.

The same complaint surfaced in a different idiom in Blantyre. A district commissioner (and my former student) pointed out, almost ruefully, that the corrupt civil servants and politicians failing the country are not strangers. They are products of the same Malawian universities that were sitting around the table. “These are the graduates,” he said. “You are all graduates.” If the higher education system is producing the people who then fail the country, what does that say about the system?

It reminded me of a doctoral researcher in Pretoria who told us about going to a major South African energy conference where the minister stood up and said, “We’re doing something we’ve never done before. We’re listening.” And then, she said, the minister walked out. And the students sat there, talking to each other.

So what does it mean for African knowledge to count? It certainly means citing African scholars. It means acting as if your own analysis is sufficient. It means noticing when you have internalized a hierarchy nobody is even enforcing anymore. And it means a serious reckoning with what universities themselves are producing.

Beyond business as usual

The most concrete area where these abstract questions come down to earth is the energy transition. A Tanzanian colleague who works on green investment opened with the right premise.

We don’t need to debate whether Africa needs energy sovereignty. That question is settled. What we need to debate is how.

His larger point was that even the so-called green economy is too often “business as usual” wrapped in new language. Solar farms displace communities. Wind projects externalize costs. Critical minerals get extracted from the eastern DRC and turn into electric cars in Berlin. The value chain remains controlled externally.

He proposed a useful test for any green investment, which he calls buying in versus buying out. If the project displaces communities, it is buying them out. If communities own a share, earn rent, or retain their land, it is buying them in. Mobile phone companies in Tanzania have figured this out, and people host towers on their property and earn a small income. There is no reason solar cannot work the same way.

His point connects to something a Zimbabwean participant said in Pretoria. When she asked residents what mattered most to them about the energy transition, the answer was not environmental sustainability. It was energy security. People want the lights to come on. They want hospitals that work.

From the South African side, an energy economist pointed out that around ninety per cent of South Africa’s coal sector is concentrated in one province, Mpumalanga. Decarbonize without a serious plan for the people who live there, and you have simply replicated the old extractive logic in a green coat.

So when African voices push back against being told to decarbonize at the same speed as Norway, this is not climate denialism. It is a serious critique of an *unjust* transition.

A Malawian colleague, Boniface Dulani, who is director of surveys at Afrobarometer, drew the comparison most starkly. Less than twenty per cent of Malawian households are connected to the electricity grid. Among those that are, supply is unreliable. So when some international donors talk about electric vehicles, the obvious question is: charged with what?

He is not against renewable energy. He is against being asked to perform a transition that wealthy countries themselves have not completed. The Afrobarometer data sharpens this. More than three quarters of Africans surveyed say wealthy nations should bear more of the cost of climate action. But around one in five also say that ordinary citizens themselves should take primary responsibility. This is not a continent that wants to be passive. It wants the burden distributed honestly.

The trouble with “homegrown”

In Addis Ababa, one of the most generative sessions focused on what people there called “homegrown solutions.” And one participant asked the right question.

What is home? Which place, which issue, are we calling home? What is the boundary?

It is an excellent question, because the word does so much work. National? Regional? Continental? Village level? Local?

The same panel offered a concrete example. The Simien Mountains National Park, in northern Ethiopia, has had decades of conflict between local communities and conservation authorities. International donors have spent serious money on the problem. The breakthrough, this past year, came when the communities themselves used their own customs and cultural practices to declare the area protected. No external cost. No major incidents.

Blantyre showed us, in much harder light, what the trouble with homegrown is.

Take the Affordable Inputs Programme — Malawi’s flagship fertilizer subsidy. Twenty years, billions of kwacha, focused almost exclusively on maize. As one participant pointed out, Malawians did not actually choose maize as the staple food. The government did, through the subsidies. Communities in northern Malawi traditionally ate cassava. And the legumes that would have done more for food security and farmer income (e.g., groundnuts, soybeans, pigeon peas) never got the same support.

Or take the Constituency Development Fund. Five billion Kwacha is now being decentralized to the constituency level. While this is a real and important transfer of power, participants described the unintended consequences. One constituency proposed three hundred small projects. Another proposed using the money to build chiefs’ houses. The decentralization is real. The mindset shift to match it is not yet there.

Then there is localization itself, which is meant to be the answer. Malawi has an NGO regulatory authority pushing international NGOs to channel funding through local organizations. So what do international NGOs do? They rebrand. One participant gave a specific example: One international NGO registered itself as a local one. Same organization. New name. Now technically a Malawian local NGO, eligible for the localization stream.

Consider also the capacity-building merry-go-round. Although some donors say local NGOs lack capacity, they often do not fund the capacity itself. They demand specific policy documents. And the names of those required documents change just often enough that small local organizations can never quite be ready. Then, having declared local organizations to lack capacity, the international NGOs hire their staff from those very local organizations. So who actually has the capacity? It is the same people. They just work in different buildings.

In my closing remarks in Blantyre I tried to push back on the entire vocabulary of “capacity building.” I have been in this business a long time and I still don’t know what it really means. We always seem to be talking about A building B’s capacity in relation to C. But seldom or never in relation to A itself. The donor will tell you to be more demanding of your government. They will not tell you to be more demanding of the donor!

Who owns the AI?

One thread ran through every conversation this year, and was the explicit focus in Mauritius last week: artificial intelligence. Almost every African government is now writing an AI strategy. Ethiopia has launched an AI University. The African Union has a draft framework. Mauritius, like many small island states, is asking how it can possibly hold its own against companies whose budgets exceed its national GDP.

My answer, which I gave in Mauritius and will repeat here, is that small states should not try to out-compute Google, Microsoft, or OpenAI. The strategy has to be asymmetric.

The most valuable AI work for places like Mauritius (or, for that matter, Malawi or Tanzania) is not training a frontier model. It is the application layer. Fine-tuning existing models for problems the country understands best. Multilingual education in Creole, French, and English. Climate adaptation for a small island. The ocean economy. Bilingual public administration. That is where competitive advantage and local sovereignty actually lie.

What I find most encouraging is that the catching-up case is no longer aspirational. It is already happening:

• CAD4TB triages tuberculosis from a chest X-ray in about five seconds, in remote settings, without internet or a radiologist.
• Ruby Health screens for anemia from a smartphone photo of a fingernail.
• Farmerline’s Darli, in Ghana, gives smallholder farmers crop-disease detection and weather updates in their own language, over WhatsApp.
• QANDA, used by around eight million students a month across Asia, gives a personalized math explanation from a phone photo.

None of these tools trains its own foundation model. What unites them is fit. They meet people where they actually are. On a basic phone, in a clinic without specialists, in a classroom without enough teachers.

However, the same technology that triages TB can also normalize mass surveillance, including predictive policing, real-time facial recognition, social scoring. These can shrink civic space long before they are used to make any specific arrest. Deepfake audio and video have been used to impersonate political leaders in elections. Large language models, combined with personality inference, can now generate politically tailored messages at scale, with no human in the loop.

There is also bias, which is not primarily a technical problem. AI systems reproduce bias because they are trained on data drawn from societies that are themselves unequal, and because their development teams remain overwhelmingly Euro-American. So my own view, in one sentence: AI will widen the gap between rich and poor countries by default, and narrow it by design.

The deciding variables (public investment in skills and connectivity, smart procurement, regional cooperation, a willingness to set rules rather than only follow them) are within reach of every government in the rooms I have described. The risk is not that the Global South will be dependent on AI in some abstract sense. The risk is that countries will accept terms of dependency they could have negotiated harder.

Five threads

At the end of the Blantyre conversation I offered five threads. They travel beyond Malawi.

• Idealism versus self-interest. Almost everyone, on every side, now operates with vested interests dressed in the language of altruism. Call out the hypocrisy, including your own.
• Evidence as afterthought. Decisions often get made on ethnic, regional, factional, or personal grounds. And evidence is wheeled out to justify what was already decided.
• Talking past each other. The minister who walks out of his own listening session. The headline that says “shift from aid to production” without really specifying how.
• Buzzwords. Resilience. End users. Capacity building. Localization. They travel well precisely because they have stopped meaning anything.
• Bargaining power. Saying “no, thank you” requires somewhere else to stand, including alliances within Africa, stronger regional institutions, South–South Cooperation, and eventually domestic resources. As one Blantyre participant put it: We are not in a social contract with our government to source aid. We are in a social contract for them to deliver services. It is our taxes funding it.

Good intentions are not the test. The test is what citizens want, and across all four rooms it came down to the same thing. They demand decent jobs, reliable electricity, a clinic that works, a school that teaches something useful, a government that does not require a bribe.

The crossroads may be the best road. But only if African citizens choose the direction. And only if it leads somewhere worth arriving at.