Artificial intelligence has been used to design fully functional new viruses that can replicate in the laboratory, in what researchers say is the first successful use of AI to create complete viral genomes.
The work, carried out by researchers at Stanford University in the US, produced 16 novel bacteriophages, a type of virus that infects bacteria. The viruses were designed to target bacteria and the researchers said they pose no threat to people.
The study marks a significant step in the use of generative AI in biology. AI tools have already been used to help design new antibiotics, but creating a viable virus from scratch is a more complex task because a complete genome must contain all the instructions needed for the virus to function and replicate inside cells.
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.”
The researchers used AI models called Evo1 and Evo2. The systems work in a way that is comparable to large language models such as ChatGPT, which predict sequences of text. In this case, the models were trained to predict sequences in genetic code rather than words.
The AI models were trained on genetic codes from viruses, bacteria, plants and people, before being refined to produce bacteriophages. These viruses infect only specific species of bacteria and are widely studied because of their potential medical and scientific uses.
The Stanford team selected the 302 most promising AI-generated designs and synthesised them in the laboratory. Of those, 16 proved able to kill E. coli bacteria.
Samuel King, a PhD student in the laboratory, said the team realised the bacteriophages were working during overnight experiments. The phages had been placed on petri dishes growing a layer of bacteria, and the researchers waited to see whether clear spots appeared where bacteria had been killed.
“We were starting to see these clear spots and it was just extremely exciting,” King said. Hie said that when the results were shared with the wider team, “the room spontaneously burst into applause”.
The findings could support future work on phage therapy, which is being investigated as a possible way to treat bacterial infections that no longer respond to antibiotics. Antimicrobial resistance has prompted renewed interest in bacteriophages because they can be highly targeted against particular bacteria.
The work also points towards broader advances in synthetic biology, where researchers aim to design biological systems beyond those already found in nature. Hie said the technology could “massively improve human health” by supporting the development of new drugs and therapies.
Independent scientists said the study was an important milestone. Prof Marc Güell, from the synthetic biology lab at Pompeu Fabra University in Spain, described it as a “very significant turning point” because, for the “first time in history, we are beginning to design biology on a computer”.
He said the approach “allows us to dream of exciting possibilities for tackling humanity's greatest challenges”, including designing phages to tackle disease, enzymes to treat genetic disorders and antibodies for use in immunotherapy.
Prof Patrick Cai, chair of synthetic genomics at the Manchester Institute of Biotechnology, also described the study as an “important milestone”. He said: “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.”
However, the research has also raised concerns about biosafety and biosecurity. Experts have warned that the same ability to design new viral genomes could be misused if applied to viruses capable of causing disease in humans, animals or plants.
In a commentary accompanying the publication in the journal Science, Dr Thomas Inglesby and Dr Moritz Hanke, from the Center for Health Security at Johns Hopkins University, wrote that the findings raise “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 Stanford researchers said they took measures to reduce risk. They excluded viruses that could infect complex organisms from the training database, focused the research on bacteriophages rather than viruses that infect people, and carried out the work in a secure laboratory.
Hie said existing safeguards already go a long way towards “ensuring that the technology is used for good”. The researchers’ approach was intended to demonstrate the potential of the technology while avoiding the creation of viruses that could infect humans.
The work does not mean AI can yet generate living organisms. Viruses are not considered alive, and the genetic scale involved in designing even the simplest living cells is much greater. The phage genome in the study is around 5,400 base pairs long, while the smallest genome of a living cell is around 500,000 base pairs. The human genome has about three billion base pairs.
Hie said it would “probably be a lot of work, but not impossible” to attempt to design some simple organisms in future, and said the team was “definitely interested in working towards” that goal.
The study adds to growing debate about how quickly AI-driven biological design is advancing and what rules may be needed to manage its risks. For supporters, the results show the potential for new tools to help treat disease. For critics, they underline the need for strict controls before such systems become more powerful and widely accessible.
