A Cambridge laboratory has run the first human trial of a vaccine whose active ingredient was designed entirely by artificial intelligence — not tweaked or analyzed by AI, but designed from scratch by it. The result is a single shot that could, in theory, protect against every known coronavirus, including strains that do not yet exist.
The trial, published in the Journal of Infection, involved 39 healthy volunteers aged 18 to 50 at clinical research facilities in Southampton and Cambridge. The vaccine was safe, caused no significant side effects and triggered immune responses against SARS-CoV-2, the original SARS virus and related bat coronaviruses that have not yet jumped to humans. A larger Phase 2 study is already planned.
For South Africa, the timing is not abstract. The country lost more than 300,000 people to excess deaths during the pandemic, saw GDP contract by 7% in 2020 and spent the better part of two years chasing variants with updated shots that often arrived after waves had broken. The idea of getting ahead of a virus rather than sprinting behind it is no longer science fiction.
Traditional vaccines are reactive. Researchers identify the circulating strain, build a vaccine around it, and hope the virus has not moved on by the time the vials reach clinics. It is why COVID-19 boosters felt like a perpetual game of catch-up — and why Professor Jonathan Heeney, who led the Cambridge research, describes the old model as “a dog chasing its tail.”
Heeney’s team took a different route. They fed an AI system the genetic codes of every Sarbeco coronavirus they could find, including bat viruses flagged by global surveillance programmes. The machine learned what these viruses shared, then designed a “super-antigen” that trains the immune system to recognize the common architecture of the entire family rather than just one member. “We’ve converted vaccine development from being reactive to being future proof,” Heeney said. “Our vaccines will continue to provide protection against viruses even as they mutate into new strains.”
The same platform is now being tested against seasonal flu, H5N1 bird flu and viral haemorrhagic fevers including Ebola. The Cambridge spinout behind it, DIOSynVax, was founded in 2017 and has been working on this architecture since before COVID-19 existed.
South Africa’s pandemic response was not without its own technological sophistication. Wits University, in partnership with York University and the Africa-Canada Artificial Intelligence and Data Modelling Consortium, built an AI-powered algorithm that predicted infection waves by analyzing mobility data, stringency indices and epidemiological parameters. The same consortium modelled vaccine prioritization strategies across the continent.
What was missing was the vaccine itself. During the Delta wave that tore through the country in mid-2021, South Africa’s infection-fatality risk in Gauteng ran to roughly 0.11%. By the time updated vaccines arrived, the damage was done. The Omicron wave that followed was milder per case but so transmissible that hospitals still buckled. A universal vaccine available before the variant emerged would have rewritten that timeline.
Professor Saul Faust, the trial’s chief investigator at the University of Southampton, said the implications go far beyond clinical research. “If we can develop and clinically advance this new class of vaccines before a virus outbreak begins, millions of lives could be saved, lockdowns avoided and the economy preserved,” he said. South Africa knows the cost of the alternative — the 2020 lockdown cost an estimated R393 billion in lost GDP, with a fiscal relief package of roughly R500 billion that did not include the businesses that never reopened or the school years that never recovered.
The immune response in the Phase 1 trial was described as “modest” — the honest assessment from the paper itself. The vaccine proved it could safely wake the immune system and target multiple coronaviruses at once, but the strength of that protection still needs to be tested at scale. Phase 2, with around 200 participants, is where that question gets answered.
Professor Andy Pollard, director of the Oxford Vaccine Group, noted that the real test is how human immune systems — shaped by years of prior infections — respond differently from laboratory animals. “It’s fascinating data, but people wouldn’t have predicted they’d be able to generate these immune responses,” he said. The unpredictability cuts both ways.
There is also the question of access. A future-proof vaccine developed in British universities and spun out through a Cambridge enterprise company will not automatically reach clinics in Soweto or Khayelitsha. South Africa’s experience with COVID-19 vaccine inequity — when wealthy nations hoarded early doses and African countries waited months — is not a distant memory. The technology is promising, even if the distribution architecture is not.
The UK’s National Institute for Health and Care Research, which funded the trial infrastructure, is calling the result a “pivotal leap forward.” Science Minister Lord Vallance framed it as “another British science success story.” Both are true, and both miss the point slightly. The success is not national — it is either global or it is not a success at all.
For South Africa, the relevant question is whether this platform, if it clears Phase 2 and beyond, can be manufactured at the scale and price that African procurement programmes can actually reach. The mRNA revolution taught the lesson that scientific breakthrough and equitable access are not the same thing.
Still, the fact that an AI can now design a vaccine antigen against viruses that do not yet exist — and that the first human trial of that design has cleared safety — is the kind of shift that redefines what pandemic preparedness means.
