Behold: an actually useful application of AI.

Researchers at the University of Cambridge have used a machine learning model to invent a new and wholly novel form of vaccine that could protect against an entire family of human coronaviruses, and even animal variants that are yet to jump over.

An initial human trial has already been completed, the results of which were published in the Journal of Infection, making it the first AI-designed vaccine to be tested in humans. The potential is dazzling: it could future-proof us against deadly virus mutations before they even occur.

“This is about making vaccines that protect us, not just from today’s viruses, but protect us from what can cause the next outbreak or disease,” study coauthor Jonathan Heeney, a Cambridge researcher, told BBC News. “This is a fundamental shift in how we prepare for pandemics.”

A traditional vaccine trains our immune system to tackle a specific virus. But these vaccines have to be constantly updated as a virus mutates over time, putting doctors on the back foot. That’s why you get new shots every flu season, and why we’re still getting new COVID-19 boosters.

AI-assisted vaccines could be a game changer. Machine learning models can scan virtually the entire corpus of genetic data collected on thousands of related viruses across the world and predict what parts of the virus are most likely to stay the same as the virus evolves, according to an expert’s analysis in The Conversation. If a vaccine focuses on these immutable virus components, it in theory could allow a vaccine to protect against an entire family of viruses — and future mutations of them, too.

And that’s that the researchers did. In the study, the Cambridge scientists focused on the sarbecovirus family, which contains both SARS and COVID, as well as strains of animal coronaviruses. After siccing their AI model on the viruses, it identified the immutable virus components, and from there, the humans narrowed down which ones were the best through careful in vivo experimentation. This allowed them to create a “super-antigen” that formed the key ingredient of the vaccine.

The results of the human trial are promising, but not definitive. It showed that the AI vaccine successfully activated the immune system to produce virus-fighting antibodies. And it was “well tolerated at all four doses with no significant safety concerns elicited,” the researchers reported in the study.

But the trial was small, with just 39 people. The immune response was “modest,” and it’s unclear how long the protection lasts.

The groundwork has been laid though, and the potential payoff would be astonishing: a potential universal vaccine. Even more encouraging is that the developed vaccine is a DNA rather than mRNA based, meaning it doesn’t require needle injections to be administered. It’s also more stable, allowing for easier transport and storage, instead of the tightly controlled, frigid conditions that mRNA vaccines require. It might even help tackle the Ebola crisis in the Democratic Republic Congo, which is being caused by a new strain that doctors don’t have a vaccine for.

“We’ve converted vaccine development from being reactive to being future proof. Our vaccines will continue to provide protection against viruses even as they mutate into new strains,” Heeney said in a statement about the work. “We’ve overcome the problem of traditional vaccines, which have limited protection. It means we can escape the constant cycle of chasing the virus variants circulating in humans and updating the vaccines to try to catch up, like a dog chasing its tail.”

More on vaccines: Ebola Expert Fears Outbreak Is Heading Into “Nightmare Scenario”