Artificial intelligence has been used to design complete viral genomes that function and replicate in a laboratory, in what researchers describe as a first for generative AI. The 16 novel viruses were designed to infect bacteria rather than humans, but the breakthrough has also prompted warnings about urgent biosafety and biosecurity risks.
BREAKING: Gates Foundation Funds First-Ever Creation of 16 Synthetic Viruses Using AI
— Nicolas Hulscher, MPH (@NicHulscher) August 7, 2026
For the first time ever, mad scientists used AI models to design entire viral genomes, physically manufacture them, and turn them into 16 new functional viruses capable of replicating.
Project… pic.twitter.com/4ZZR4hk8ts
The achievement could open new possibilities in medical research, particularly in the search for treatments for antibiotic-resistant infections. But it also marks a significant escalation in what AI systems can design, moving beyond individual genes and proteins to complete genetic systems capable of functioning inside living cells.
'This is a next step in the complexity that is designable by generative AI,' Brian Hie, an assistant professor at Stanford University, told the BBC. He described it as the first time generative AI had been used to design a complete genome capable of replicating and performing other functions inside living cells.
How AI Designed The Viruses
The research used AI models called Evo1 and Evo2, which were trained to recognise patterns in genetic sequences. The approach is broadly comparable to the way language models predict sequences of words, although the biological systems involved are considerably more complex.
The researchers used the models to generate designs for bacteriophages, viruses that infect bacteria. Their work focused on the bacteriophage ΦX174, which infects E. coli and has a relatively small genome.
Of the hundreds of AI-generated designs assessed by the researchers, 302 of the most promising candidates were synthesised and tested in the laboratory. Sixteen successfully infected and killed E. coli bacteria. Samuel King, a PhD student involved in the research, said the team realised the designs were working when they began seeing clear areas where bacteria had been destroyed.
'We were starting to see these clear spots, and it was just extremely exciting,' King said. Hie recalled that the room later 'spontaneously burst into applause.' The result demonstrated that an AI-generated genetic sequence could do more than simply resemble a biological genome. It could produce a functioning viral system.
Potential For Medical Research
The researchers believe the technology could eventually contribute to new treatments for bacterial infections, including infections that have become resistant to conventional antibiotics. In laboratory experiments, a combination of the AI-designed phages was also able to suppress E. coli that had developed resistance to naturally occurring phages.
Hie said the technology could 'massively improve human health' by helping researchers develop new drugs and therapies. Prof. Marc Güell of Pompeu Fabra University in Spain described the study as a 'very significant turning point.'
Prof. Patrick Cai, chair of synthetic genomics at the Manchester Institute of Biotechnology, called the research an 'important milestone.'
'The significance extends far beyond phages—it suggests that genome language models are beginning to learn the design principles encoded by evolution, opening the door to AI-assisted genome writing,' Cai said. But the same capability that could help researchers tackle disease is also raising difficult questions about how the technology could be misused.
Scientists Warn Of Risks
In a commentary accompanying the study in Science, Dr Thomas Inglesby and Dr Moritz Hanke of the Johns Hopkins Center for Health Security warned of 'urgent biosafety and biosecurity questions.'
They argued that the issue was no longer whether generative AI could be used for viral genome design, but whether the technology could be developed without creating opportunities for serious harm. They also cautioned against pursuing viruses with the potential to cause disease.
The researchers took precautions to reduce those risks, including restricting the biological material used to train the models and focusing on bacteriophages rather than viruses that infect humans. The wider safety challenge, however, extends beyond the AI models themselves. Experts have raised questions about DNA synthesis screening, laboratory safeguards, and whether existing systems can reliably identify potentially dangerous genetic material before it is produced.
From Phages To More Complex Biology
The viruses created in the study are a long way from being anything resembling a human pathogen. Their genomes are relatively small compared with those of cellular organisms, and the researchers did not create viruses capable of infecting people.
Hie nevertheless acknowledged that designing increasingly complex biological systems could eventually become possible. He said creating simple organisms would probably require considerable additional work but was not necessarily impossible.
That remains a significant distance from the achievement reported in the study. For now, the breakthrough is narrower but still important. AI has successfully generated complete viral genomes that function inside living cells.
That could help scientists develop new ways to fight bacterial disease. It also establishes a new frontier in biological design—one where the question is no longer simply what AI can predict, but what it can create.
The challenge now is ensuring that safety, oversight, and regulation keep pace with a technology whose capabilities are advancing faster than many existing safeguards.
