For the first time, an artificial intelligence-designed antigen has been tested in humans, after scientists at the University of Cambridge created a vaccine intended to work against all types of coronavirus — including every variant of Covid-19 and animal viruses with the potential to trigger future pandemics.
AI Designs ‘Super-Antigen’ from Viral Genetic Codes
The Cambridge team began by gathering genetic codes from multiple coronaviruses identified through surveillance programmes that monitor potential viral threats. An AI system analysed this data and designed what the researchers call a “super-antigen” — a component capable of training the immune system to recognise entire virus families rather than individual strains.
The super-antigen targets conserved features within a virus family, elements that the pathogen cannot easily change. Professor Jonathan Heeney, from the University of Cambridge, described the approach as a “fundamental shift in how we prepare for pandemics”. He said: “We’re always behind,” explaining that the goal is to “get ahead of the curve” to protect against new outbreaks before they emerge.
Traditional vaccines rely on current strains of virus to function, leaving the world in a reactive posture. The AI-designed antigen, by contrast, is intended to train the immune system to defend against mutated versions of known viruses and novel infections that jump from animals to humans, such as zoonotic coronaviruses. Professor Heeney said the technology was “surprising all of us” and described it as “amazing what we can do with it for the good of humanity”.
Initial safety trials involved 39 participants and produced what researchers described as “modest” immune responses — yet the findings, published in the Journal of Infection, have generated considerable excitement among scientists. Professor Marian Knight, of the National Institute for Health and Care Research, called the trial a “significant milestone” that could lead to broader and more durable protection against viral diseases.
The vaccine was administered using a needle-free microfluid jet, which delivers vaccine blueprints directly into skin cells, offering an alternative for those with a fear of injections.
Universal Protection Against Coronavirus Families
The vaccine is designed to provide protection against all types of coronaviruses, including every variant of SARS-CoV-2 that causes Covid-19, as well as animal coronaviruses that could spark future outbreaks. Because the AI system targeted genetic sequences common to many coronaviruses — rather than those unique to a single strain — the resulting antigen should offer what researchers describe as “future-proofed” protection, effective even as the virus mutates.
Professor Saul Faust, who conducted some of the trials at the University of Southampton, said the AI design “definitely has potential” and was “really exciting”, particularly for designing vaccines against rapidly changing viruses with pandemic potential. He told the BBC: “What’s really interesting is the technology is an awful lot better at designing vaccines for potential pandemics when viruses are changing.” He noted that the current reactive vaccine system struggles to keep pace with evolving viruses.
A second study with approximately 200 volunteers is already underway to provide deeper insight into how effectively the vaccine trains the immune system. Professor Andy Pollard, director of the Oxford Vaccine Group, said artificial intelligence would be a “game-changer” for vaccine research, capable of predicting immune responses and speeding up development. He pointed out that while animal research has been compelling, human trials are the crucial next step because of differences in immune systems between species.
Future Potential: From Flu to Ebola
The Cambridge team is already applying the same AI technology beyond coronaviruses. They are conducting animal research on universal seasonal flu vaccines that would eliminate the need for annual updates — a long-sought goal in influenza prevention. They are also working on an H5N1 bird flu vaccine, in case the virus currently devastating bird populations spreads to humans.
Research into vaccines for viral haemorrhagic fevers, including Ebola species, is also underway. The current outbreak in the Democratic Republic of Congo involves the Bundibugyo virus, a species for which no licensed vaccine exists. The existing ERVEBO® vaccine is approved for certain at-risk individuals only against Orthoebolavirus zairense and is not effective against Bundibugyo. There is also no vaccine for Marburg virus disease. The Cambridge team’s AI approach could potentially offer broad protection against the entire Ebola group of viruses.
Separately, Professor Pollard is leading a £118 million initiative with the Ellison Institute of Technology called CoI-AI (Correlates of Immunity-Artificial Intelligence). The programme combines Oxford’s expertise in immunology and human challenge studies with AI to better understand how the body fights infection and how vaccines protect us, with a focus on pathogens that have so far evaded prevention — including Streptococcus pneumoniae, Staphylococcus aureus, and E. coli.
Concerns Over AI and Bioweapons
Yet the same technology that promises to revolutionise vaccine development has prompted warnings from within the AI industry itself. Several major AI firms — including Google DeepMind, OpenAI, and Anthropic — have written to the US Congress urging the adoption of new laws to make it harder for “bad actors” to develop biological weapons using similar tools. The signatories — OpenAI’s Sam Altman, Google’s Demis Hassabis, and Anthropic’s Dario Amodei — called for legislation requiring companies that sell synthetic DNA and RNA to screen customers and orders to prevent the misuse of genetic material. Experts argue that the danger lies not just in the AI’s output but in the connection of digital knowledge to physical materials, and they advocate for regulating the conversion points where AI capabilities translate into real-world consequences, focusing on supply chain safeguards, laboratory practices, and customer verification.