Researchers at the University of Cambridge have developed an AI-assisted vaccine technology designed to provide protection against entire virus families rather than individual virus strains. The approach could fundamentally change vaccine development by creating broader immunity against both existing and emerging viruses, potentially helping to prevent future pandemics.
Unlike conventional vaccines, which are typically developed in response to circulating virus variants, the new technology uses artificial intelligence to identify shared characteristics across multiple viruses within the same family. The researchers describe the approach as a shift from reacting to outbreaks toward anticipating them. The first application of the technology is a universal vaccine targeting sarbecoviruses, the coronavirus family that includes both SARS-CoV and SARS-CoV-2. Early clinical results suggest the vaccine is safe, paving the way for larger clinical studies.
AI searches for the virus 'master key'
According to Professor Jonathan Heeney of the University of Cambridge, traditional vaccines are constantly trying to catch up with viral evolution. By the time a vaccine has been developed, manufactured and distributed, new variants may already have emerged. To overcome this challenge, the Cambridge team used artificial intelligence to analyse large datasets containing information on multiple viruses. Rather than focusing on the unique characteristics of individual strains, the AI identified stable viral regions shared across entire virus families that are recognized by the human immune system.
Heeney compares the technology to a "master key" capable of opening every apartment in a building instead of carrying separate keys for each individual door. By targeting conserved viral structures, the researchers hope to generate immune responses that remain effective even as viruses continue to evolve. The technology also benefits from advances in modern AI, enabling researchers to analyse increasingly large and complex biological datasets and accelerate vaccine design.
Lessons learned from Ebola
The project originated after the devastating Ebola outbreak in West Africa between 2013 and 2016. At the time, Heeney was working in the region and witnessed firsthand how valuable time was lost before the virus was correctly identified and vaccine development could begin. According to the World Health Organization, the outbreak ultimately claimed approximately 11,300 lives. During the first months of the epidemic, the virus spread rapidly across Guinea, Sierra Leone and Liberia before effective countermeasures became available.
The experience convinced the researchers that vaccine development needed a fundamentally different strategy. As population growth, global travel and increasing human encroachment into wildlife habitats continue to drive the emergence of new zoonotic diseases, the need for broader and faster vaccine platforms is becoming increasingly urgent.
Promising early results
The AI-designed sarbecovirus vaccine, developed by Cambridge researchers together with biotechnology company DIOSynVax, has completed an initial clinical study involving 39 healthy volunteers. The trial, sponsored by University Hospital Southampton and published in the Journal of Infection, reported no significant safety concerns. The vaccine will now advance to larger clinical trials to further evaluate its safety and effectiveness. Meanwhile, the research team is continuing to expand its AI platform using the latest generative AI technologies to accelerate vaccine discovery and optimise future vaccine candidates.
Although additional clinical validation will be required, the researchers believe the platform could mark the beginning of a new generation of vaccines capable of providing broad protection against entire virus families. If successful, the approach could significantly improve global preparedness for future infectious disease outbreaks and reduce the time needed to respond to emerging pandemic threats.
Pandemic preparedness
Last year, scientists from the University of Oxford and partners across academia, industry and public policy outlined how AI could significantly strengthen global pandemic preparedness. The study highlights AI's potential to improve disease surveillance, predict outbreaks, identify emerging variants, accelerate vaccine development and support outbreak detection by combining population-level data with individual data from sources such as wearable devices. Researchers also believe AI could help identify high-risk regions, predict pathogen characteristics and make advanced scientific insights more accessible to healthcare professionals worldwide.
However, the authors stress that AI alone is not sufficient. They call for greater international collaboration, transparent datasets and AI models, and stronger governance around safety, accountability and ethics. Concerns remain about data quality, representativeness, limited access to AI models and the risks of relying on "black-box" systems. According to the researchers, combining AI with continuous surveillance, high-quality data and human expertise will be essential to better prepare for future pandemics.
