Cambridge: A University of Cambridge-led research team has reported encouraging early results from a human trial of an experimental AI-designed vaccine aimed at providing broader protection against coronaviruses that could pose future pandemic risks.
They worked with DIOSynVax, a Cambridge spin-out, to design the vaccine. It’s meant to target the bigger sarbecovirus family, which includes SARS-CoV-2 (the virus behind COVID-19), SARS-CoV-1, and a range of bat coronaviruses. Some of these bat viruses haven’t infected humans yet, but experts worry they could jump to people down the line.
For this vaccine, the scientists didn’t just aim at one strain. Instead, they let AI and computer simulations sift through the genetic data of all these sarbecoviruses, drawing from surveillance programs around the world. The technology searched for genetic features the viruses share. With those common features, the team crafted what they call a “super-antigen.” Basically, it’s a piece of the vaccine designed to train the immune system to spot and fight off a broader variety of related viruses, instead of locking onto a single target.
They ran an early trial with 39 healthy volunteers. No major safety issues came up—the vaccine was safe and generally well tolerated. And the volunteers developed immune responses not just against COVID-19, but also SARS and even a few related bat viruses. These results are encouraging. They show that using AI to help design vaccines really can work. But the immune responses in this first trial were modest, and scientists say we need more research before this vaccine is anywhere close to being widely used.
One detail that stands out: the vaccine was delivered as a DNA shot, using a needle-free jet system that pushes the vaccine into the skin. This kind of delivery could make vaccinations easier on a large scale, especially in places where traditional injections are tricky. This matters because most vaccines are always playing catch-up—when viruses change, the shots need updating. It’s a familiar story with COVID boosters and flu shots every year.
The Cambridge approach flips that around, aiming to get ahead of potential new threats by designing vaccines ready for whole virus families, not just last year’s strains. Scientists hope these broad-spectrum vaccines can buy public health systems more time when a new virus appears—maybe catching an outbreak before it gets out of hand. If later trials show even stronger and longer-lasting protection, using AI for vaccine design could seriously speed up the process of identifying promising vaccine candidates.
Still, this research is just beginning. There’s no real-world proof yet that the vaccine actually stops people from getting sick. We don’t know how strong or lasting its protection might be, whether people will need boosters, or how it’ll perform in bigger, more diverse groups. A larger Phase 2 trial is on the way to look for stronger, broader immune responses and see how well this strategy holds up.
Stepping back, this study marks a real milestone for AI in vaccine development. No, it doesn’t mean a universal coronavirus vaccine is in our hands tomorrow. But it does show AI can help scientists focus on vaccine targets a lot faster than old-school methods. And for global health, one lesson is obvious: getting ready for future pandemics isn’t just about acting quickly after an outbreak. It’s about designing smarter vaccines before the next threat shows up.
