By Tanatsiwa Dambuza
It is a well-known fact that most of Africa’s giving is never officially counted. Burial societies pool money to put the deceased to rest each month. Savings rounds put robbed traders back into business. Diaspora families wire funds home for boreholes and school fees. Neighbours feed neighbours, and traders extend credit across borders on nothing but trust.
Despite the informal nature of these giving practices, the digital lives of these same communities, including their messages, mobile payments and marketplace listings, are harvested at an industrial scale and turned into profit on other continents. This story sets those two facts side by side, like two ledgers. One ledger, kept abroad, records everything Africans do online and misses nothing. The other ledger, which would record what Africans give, has never been opened. The story examines who keeps the first ledger and profits from it, why the second does not exist and what the African Continental Free Trade Area’s (AfCFTA) new digital trade rules could change. Finally, it will explore what decolonising data could mean for a continent whose generosity is enormous, yet invisible.
The Ledger They Keep
Scholars call the first ledger data colonialism, a term made prominent by the researchers Nick Couldry and Ulises Mejias. African thinkers such as Everisto Benyera have gone further, arguing that the fourth industrial revolution amounts to a recolonisation of Africa by digital means. The idea is simpler than the vocabulary. In the old colonial economy, the continent’s minerals left raw and came back expensive. In the new one, it’s data that leaves and returns in the same manner.
The entries in this ledger are precise. Africa holds 18 percent of the world’s population but supplies less than 4 percent of the data used to train the most powerful artificial intelligence systems. It is why those systems so often misread African realities, from credit scoring to medical screening. The continent hosts about 0.6 percent of global data centre capacity and foreign servers handle an estimated 80 percent of its internet traffic, which means African data lives under foreign law before it lives under African law. Fewer than 0.1 percent of online resources exist in African languages, against the much larger 53 percent in English.
When Africans appear inside this economy, they appear as cheap labour. A TIME investigation found Nairobi based workers paid between 1.32 and 2 dollars an hour to label the internet’s most disturbing content so that a chatbot would be safe for everyone else, while documents reviewed by60 Minutes showed OpenAI had agreed to pay Sama, the San Francisco outsourcing firm running the Nairobi operation, 12.50 dollars an hour per worker, six to nine times what the workers themselves received. Digital platforms now extract not only African data, but attention and nerve. The damage stayed in Nairobi while the product went to market in San Francisco.
When Africans appear inside this economy, they appear as cheap labour. A TIME investigation found Nairobi-based workers paid between 1.32 and 2 dollars an hour to label the internet’s most disturbing content so that a chatbot would be safe for everyone else, while documents reviewed by
60 Minutes showed OpenAI had agreed to pay Sama, the San Francisco outsourcing firm running the Nairobi operation, 12.50 dollars an hour per worker, six to nine times what the workers themselves received. Digital platforms now extract not only African data, but attention and nerve. The damage stayed in Nairobi while the product went to market in San Francisco.
The Ledger Nobody Keeps
The second ledger is the one that records African giving, except its pages are blank.
Nine years on the Harare-to-Lusaka route have shaped Rudo’s (27) month: contributing to a savings round to help a robbed colleague recover, sharing the funeral costs for a stranger, and ferrying parcels six hundred kilometres for neighbours at no charge.
“The phone is my shop, my bank, my everything,” she said. “Every dollar I spend, it knows. But what I give? That one it does not see. There is no button for it.” Every transfer Rudo makes generates a message to her handset and a line in a database. What it does not record is what the payment was for. Twenty dollars into a funeral collection and twenty dollars for stock look identical in the ledger: sender, recipient, amount, timestamp. The system captures the movement and discards the meaning. That is the button Rudo means. Not a missing record, a missing category.
Counting it would carry its own risk. Much African giving is deliberately quiet, the school fees settled without announcement, the funeral contribution made so as not to embarrass the bereaved. Anonymity protects the recipient from obligation and the giver from the suspicion of display. But the choice is not between being seen and being left alone. Rudo’s phone already reports her to companies she has never heard of, and her giving is already visible to the people it concerns. What is missing is not privacy but control: who holds the record, and who decides what it is for. Counting on terms the community sets means aggregation rather than names, and a record held not by the mobile operator but by an African institution answerable to the people in it, a community trust, or a public research body rather than a platform.
Her instinct holds at the highest levels of African philanthropy. Naledi Dube, a programme officer at a pan-African foundation, spent two years trying to measure community giving across several countries. The project failed, and she is precise about why.
“We could tell you, to the dollar, what those communities spend. Airtime, groceries, and transport—it is all in somebody’s data,” she said. “We could not tell you what they give. The giving is in there, but no platform has ever had a reason to see it. So the world’s picture of giving gets built from tax records and foundation databases, and a continent that gives every single day gets recorded as a continent that receives.”
The scale of the blank pages is suggested by the pipes the giving flows through. Mobile money moved about 1.43 trillion dollars across Africa in 2025, roughly two-thirds of the world total, and somewhere inside that torrent sits the school fees paid for a sister’s child, the borehole collections, and the funeral transfers. No dataset anywhere separates the gift from the purchase.
Some giving never reaches the network at all. Chiedza (23), a Harare seller who trades clothing through WhatsApp and Facebook, hands stock every month to two widows in her community who resell it and keep the proceeds. No money moves, so nothing is logged, categorised or corrected. Where Rudo’s giving is recorded but stripped of meaning, Chiedza’s leaves no trace to interpret at all.
Two invisibilities, then, and they are not a coincidence. They are the same arrangement seen from opposite sides. African giving remains invisible to the data economy because it occurs in areas of life that platforms cannot price. African data is invisible to the givers who generate it because the systems that hold it were never built to serve them.
Balancing the books
Africa now has a legal instrument with which to reopen the accounts. In February 2024 the African Union adopted the AfCFTA Protocol on Digital Trade, the continent’s first comprehensive digital trade agreement, whose Articles 20 to 23 set the framework for cross-border data transfers, data protection and data governance. In February 2025, the AU adopted eight annexes, reaching into emerging technologies, including artificial intelligence. African states have committed 60 billion dollars to continental AI ambitions, and the AfCFTA Secretariat says the continent needs more than 700 data centres by 2035.
The instrument may also do the opposite of what is hoped of it. It is a trade agreement before it is anything else, and a trade agreement exists to reduce friction. Harmonising the rules under which data crosses fifty-four borders makes African data easier to move. The parties best placed to use a frictionless continental data pool are those with the capacity to store and process it, which sits overwhelmingly outside the continent. The protections in Articles 20 to 23 are real, but they are permissive and they depend on national implementation that has largely not happened. Harmonisation arrives first. Enforcement arrives later, if at all. Tanaka Ndongera, a digital trade policy analyst who has followed the protocol negotiations closely, calls the paper the easy part. “The protocol gives Africa the rules. It does not decide who benefits from them,” he said. “We have made it lawful and simple to move African data across African borders at a moment when almost nobody here has the computing power to do anything with it. That is an efficient pipeline with the far end handed to whoever owns the servers. The window in which these rules could be shaped around ordinary people is the same window in which AI is racing to lock in the old arrangement.”
This is where decoloniality stops being a seminar word and becomes a work plan, because decolonising data was never only about policing the extractors better. It is about Africans owning the means of seeing themselves. Dr. Levious Chiukira, a multidisciplinary researcher at Midlands State University, explains what that ownership would look like. Continental datasets of community giving held by African institutions and governed by the communities they serve; community data trusts, so the value people generate is negotiated collectively rather than surrendered one handset at a time; and artificial intelligence built in African languages, because a model that cannot read Shona or Zulu cannot see the generosity spoken in them.
“An AI trained on Africa’s reality would know this continent runs on reciprocity,” he said. “The machines have read our messages, and they still do not know that about us. We keep asking for a better price for our data. The larger prize is building systems that work the way our societies actually work, because nobody else is going to build that for us.”
His last point is the one most easily missed. A demand for a fairer share accepts the shape of the thing being shared and leaves intact the old story of Africa as a source of inputs; minerals once and data now. The economic life described here runs on something the dominant models do not encode: obligation that is not priced and credit extended without paper, because those models were built to broker transactions between strangers. That is not a quaint survival to be photographed and filed away. It is a working logic indigenous to Africans, and it is the one thing African systems could be built on that nobody else could reason to build.
Building those eyes is the philanthropic project of this decade, and the burden falls on African philanthropy itself. Existing digital platforms profit from blindness and governments move at the speed of ratification. But the foundations and funders who speak constantly of African generosity have never financed the instruments that would prove it exists. The legal framework, the mobile money rails and the AI moment have arrived at once, and that failure has run out of excuses. A sector that cannot see its givers cannot claim to serve them.
Until someone opens the second ledger, the first one keeps compounding. Africa’s data will go on enriching people who have never set foot on the continent, and Africa’s givers— the traders, the savings rounds, the church women, and the diaspora— families will go on carrying the largest philanthropic system in the world, invisibly, one uncounted gift at a time.