AI just created a real virus in the lab that no one has heard of before
AI has just made another breakthrough – it has created new viruses. Now, a new virus may sound alarming, but researchers at Stanford University and
AI has just made another breakthrough – it has created new viruses. Now, a new virus may sound alarming, but researchers at Stanford University and the Arc Institute have used AI to design 16 brand-new viruses that can infect bacteria but not harm humans. The study marks what researchers described as the first time AI has been used to design whole viral genomes that went on to function in the laboratory. Read Full Story This could be a "very significant turning point" in science that may allow us to enter a new era for treating disease. At the same time, there are safety concerns over AI-designed viruses. AI creates virus, humans don’t need to worry The study produced 16 new viruses that were able to infect and reproduce inside E. coli bacteria. Scientists said the project was deliberately limited to bacteriophages and carried out in a secure laboratory. The researchers used genome language models called Evo 1 and Evo 2. These systems work in a way similar to text-generating AI tools, but instead of learning language from books or websites, they learn patterns in DNA sequences.
The models were trained on vast amounts of genetic data, including sequences from viruses, bacteria, plants and people, and then refined to focus on a narrow family of bacteriophages related to Phi After training, the AI generated a very large number of possible viral genomes. Scientists then screened the results and selected around 300 of the most promising designs for testing. The DNA was synthesised in the laboratory and introduced into bacteria. Of these candidates, 16 went on to produce viable viruses that could replicate inside bacterial cells. According to the study, some of the synthetic phages performed better than Phi Next step for AI AI has managed to progress rapidly in recent months. From simple generative text to coding for hours on its own. But creating complete genomes on its own is being seen as the next step in generative AI. Brian Hie, assistant professor at Stanford University, told the BBC, "This is a next step in the complexity that's designable by generative AI, this is the first time generative AI has been used to design a complete genome, it's something that can replicate and have other functions inside cells this was new territory for us." Hie said the technology could "massively improve human health" by helping create new drugs and therapies.
Researchers said the work may also help efforts to develop phage therapy, a long-studied approach in which viruses are used to target harmful bacteria, including strains that no longer respond to antibiotics. One of the successful viruses carried an "evolutionarily distant" genetic feature, suggesting the AI produced a design that may have taken natural evolution a very long time to arrive at. This is not the first time AI has made a breakthrough in biology. Google DeepMind’s AlphaFold can successfully predict the 3D shape of proteins from their amino acid chains. Google DeepMind’s Demis Hassabis and John Jumper (now at Anthropic) won the Nobel Prize for creating AlphaFold in 2024. A cause for alarm? Creating a virus that does not harm humans, plants or animals, sounds ideal. But the study has sparked fears over the potential AI has unlocked in genome creation. In a commentary published alongside the paper, Dr Thomas Inglesby and Dr Moritz Hanke from the Center for Health Security at Johns Hopkins University wrote that the findings raised "urgent biosafety and biosecurity questions.” They said the issue was no longer whether generative viral genome design would exist, but whether it could be used without "enabling serious harm.” They added that new viruses with the potential to cause disease "should not be pursued.” The researchers said their work did not involve viruses that infect people and does not show that AI can currently design human pathogens.
The BriefWire