A University of Cambridge team has trialled in 39 healthy adults the first vaccine antigen designed entirely by artificial intelligence, testing a candidate meant to protect against all coronaviruses, including SARS‑CoV‑2 and related bat viruses.
The small first-in-human study, run at NIHR clinical research facilities in Southampton and Cambridge and sponsored by University Hospital Southampton NHS Foundation Trust, found the AI‑designed antigen was safe and caused no significant side‑effects in volunteers aged 18 to 50. Results published in the Journal of Infection also showed the vaccine triggered immune responses to SARS‑CoV‑2, SARS and other Sarbeco coronaviruses.
Researchers built the antigen from genetic sequence data collected by global surveillance programmes. Their artificial intelligence sifted known coronavirus genomes and designed a “super‑antigen” intended to present parts of the virus that change less as viruses evolve. The aim is a type of vaccine that remains protective as pathogens mutate, rather than one based on a single circulating strain that must be regularly updated.
That goal is the trial’s central friction. While volunteers did mount measurable responses to multiple coronaviruses, the study’s authors describe the immune‑system impact as modest. The finding establishes safety but leaves open whether the antibody and cellular responses recorded will translate into strong, durable protection against infection or disease.
The team frames the work as a change in strategy for vaccines. Cambridge researchers say designing antigens by AI can aim coverage at entire virus families and animal viruses with spillover risk, potentially reducing the need to chase newly circulating variants. The same computational approach is already being used in parallel efforts to develop candidate vaccines for influenza and Ebola.
Researchers involved argue this method could move vaccine development from a reactive cycle of updating shots for each new variant to a more forward‑looking model. Jonathan Heeney, part of the Cambridge group, has emphasised the intention to get ahead of future outbreaks and to produce vaccines that protect against what could cause the next disease event rather than only today’s viruses. Independent academics overseeing the trial called the work promising, saying it has real potential and is an exciting step for technologies that design vaccines for changing pathogens.
What happens next is explicit and immediate: a second study of around 200 participants will measure how well the antigen trains the immune system. That larger trial is intended to quantify the strength and breadth of responses and to see whether the modest signals seen so far can be amplified, or whether different formulations or adjuvants are needed to produce protection on a scale useful for public health.
The unanswered, consequential question is whether a vaccine component conceived entirely by AI can move beyond safety and modest immunogenicity to prevent infection and disease when tested in larger, more diverse populations. If it does, the approach could change how vaccines are built for future pandemics; if it does not, the field will need to refine the computational designs or combine them with other strategies to achieve broader, stronger protection.
