Recently, in the journal Science, researchers at Stanford University and the Arc Institute reported something that had never been done before. An artificial intelligence designed a complete, functioning genome from scratch. The models, trained not on human language but on trillions of nucleotides of genetic code, generated thousands of candidate viral genomes. Nearly three hundred were chemically synthesised. Sixteen came alive: bacteriophages, viruses that prey on bacteria, resembling nothing found in nature. The most consequential result was almost buried in the coverage. A cocktail of these designed phages rapidly destroyed bacteria that had already evolved resistance to a natural phage. That is not a laboratory curiosity. That is a proof of concept for an entirely new class of medicine against drug-resistant infection. Nowhere on earth should that sentence land harder than here. The burden is ours The Global Research on Antimicrobial Resistance study, published in The Lancet, found that in 2019 bacterial resistance was directly responsible for roughly 1.27 million deaths worldwide and associated with nearly five million more. The highest burden of any region on earth was in sub-Saharan Africa, at approximately 24 deaths per 100,000 people, roughly double the rate in high-income countries. A 2024 update projects some 39 million cumulative deaths from resistant infection between 2025 and 2050, with the sharpest increases in the very regions least equipped to absorb them. Tuberculosis, which this continent knows well, tells the same story. The World Health Organisation counts roughly 400,000 cases of multidrug-resistant TB each year, and fewer than half of those people are ever diagnosed and started on effective treatment. We are not waiting for a designed pathogen. We are already losing to ordinary ones that have stopped answering to ordinary drugs. So when commentators in London and Washington describe AI-designed viruses primarily as a biosecurity threat, they are not wrong. But they are speaking from a position of relative safety. The asymmetry deserves naming. Wealthy nations fear what this technology might one day release. We are dying from what it might one day cure. The dual-use problem, stated honestly None of these dissolves the danger. The same architecture that produced a therapeutic phage could, with a different training set, be pointed at pathogens that infect people. The researchers' principal safeguard was to exclude from training any virus capable of infecting humans, animals, plants, or fungi. That is a responsible choice. It is also a procedural one, reversible by any laboratory with a broader dataset and fewer scruples. This is the classic dual-use dilemma, but with two new features. The first is speed: design cycles that once took years now take days. The second is diffusion: the barrier to entry is shifting from tacit laboratory craft, which is scarce, to model access and synthesis capacity, which is becoming abundant. Governance built for the era of the specialised wet lab is not built for this. Where Africa's real risk lies Here is the argument I want to press on readers of this newspaper. Africa's greatest exposure in the age of designed life is not that a synthetic pathogen will be aimed at us. It is that the rules governing this technology will be written without us, and that we will once again receive the risks early and the benefits late. The evidence for that pattern is not speculative. Africa carries roughly a quarter of the global disease burden and hosts something on the order of two percent of the world’s clinical trials. The continent has historically manufactured less than one percent of the vaccines it administers. During the COVID-19 pandemic, Dr. Tedros Adhanom Ghebreyesus warned that the world stood on the brink of a “catastrophic moral failure” in vaccine equity. He was right, and the failure was not averted. Dr. John Nkengasong, who led the Africa Centres for Disease Control and Prevention through that period, drew the correct conclusion and gave it a name: a New Public Health Order, built on African manufacturing, African regulatory authority, and African science. Chinua Achebe was fond of quoting an Igbo proverb: until the lions have their own historians, the history of the hunt will always glorify the hunter. Bioethics has its own version of that problem. The ethics of emergent biotechnology is currently being written almost entirely by institutions in countries that will not bear the first cost of getting it wrong. What Ubuntu asks of a synthetic biology There is a distinctly African contribution to make here, and it is not decorative. John Mbiti’s formulation, “I am because we are,” is not sentiment. It is a claim about where moral standing comes from, and it generates obligations that the dominant Western bioethical frame, organised around individual autonomy, tends to under-specify. Applied to designed life, it yields three demands. First, that consent be understood communally as well as individually: populations, not only patients, have a stake in what is released into their environment. Second, that benefit-sharing be a condition of research rather than a courtesy extended afterward. Third, that governance be participatory. A New Public Health Order requires a corresponding New Bioethical Order, one in which African ethicists, regulators, and scientists sit on the drafting committees and not merely on the consultation lists. Paul Farmer put the underlying principle plainly: the idea that some lives matter less is the root of all that is wrong with the world. That idea is rarely stated. It is enacted, quietly, through the sequencing of who gets access to what, and when. Rwanda's particular obligation This country is unusually well positioned to act rather than comment. The African Union has set a target of manufacturing sixty percent of the continent’s vaccines locally by 2040. Rwanda has already committed institutional capital to biomanufacturing and regulatory strengthening, and has hosted continental health governance with credibility. That combination, sovereign capital plus regulatory ambition plus political seriousness, is rare anywhere in the world. Four things follow. Rwanda and its partners should build genuine domestic capacity in AI-enabled biology rather than importing finished conclusions. The African Medicines Agency should develop an anticipatory framework for AI-designed biological agents now, before the first application arrives on a desk. African bioethicists should be funded to publish, not merely to review. And the continent should insist that any therapeutic emerging from this technology carries an access commitment written into its financing from the first term sheet. Julius Nyerere said we must run while others walk. Steve Biko warned that the most potent weapon in the hands of the oppressor is the mind of the oppressed. Both cautions apply. The technology described in Science last week is neither salvation nor apocalypse. It is capability. The question of who holds it, who governs it, and who benefits from it is not a technical question at all. It is a moral and institutional one, and it is being decided now. We should not receive that decision. We should help make it. Michael Fairbanks is Co-Founder and Executive Chairman of Akagera Medicines, a clinical-stage biopharmaceutical company headquartered in Kigali, and a Fellow at Harvard Medical School's Center for Bioethics and the Harvard Weatherhead Center for International Affairs.
