WASHINGTON, D.C. — Scientists in the United States have used artificial intelligence to design entirely new viruses capable of functioning and replicating in the laboratory, marking a major scientific milestone that researchers say could transform medicine while raising fresh concerns over biosafety and biosecurity.
The breakthrough, led by researchers at Stanford University, marks the first time complete viral genomes have been designed using generative AI and successfully shown to work in laboratory experiments.
The study, published in the journal Science, involved designing 16 previously unknown viruses known as bacteriophages, or phages, which exclusively infect bacteria and pose no threat to humans.
Researchers say the achievement could accelerate the development of new treatments for antibiotic-resistant infections, while independent experts caution that the same technology could eventually be misused to engineer dangerous pathogens if adequate safeguards are not maintained.
AI writes the ‘language of life’
Unlike conventional artificial intelligence systems that generate text or images, the models used in the study—known as Evo 1 and Evo 2—were trained to understand genetic sequences, enabling them to predict and generate DNA rather than words.
The researchers trained the AI on millions of genetic sequences from viruses, bacteria, plants and animals before refining it specifically to design bacteriophages capable of targeting Escherichia coli (E. coli) bacteria.
From hundreds of AI-generated viral genomes, scientists selected 302 candidates for laboratory testing.
Sixteen of those synthetic viruses successfully infected and destroyed E. coli, confirming that the AI-designed genomes were fully functional.
“This is a next step in the complexity that’s designable by generative AI,” said Brian Hie, an assistant professor at Stanford University and one of the study’s lead researchers.
“It is the first time generative AI has been used to design a complete genome—something that can replicate and have other functions inside cells.”
A breakthrough against antibiotic resistance
Scientists believe the research could provide a powerful new tool in the fight against antimicrobial resistance, one of the world’s fastest-growing public health threats.
Bacteriophages naturally prey on bacteria and have long been explored as an alternative to antibiotics, particularly against infections that no longer respond to conventional drugs.
Artificial intelligence could significantly shorten the time required to develop tailor-made phages capable of targeting specific bacterial strains, potentially opening new treatment options for difficult-to-treat infections.
The researchers described the moment they confirmed the AI-designed viruses were working as one of the most exciting points of the project.
Doctoral researcher Samuel King said laboratory tests revealed clear zones on petri dishes where the engineered phages had destroyed bacterial colonies.
“When the results came through, the room spontaneously burst into applause,” Hie recalled.
Safety concerns emerge
Despite the medical promise, the research has also reignited debate over the risks of rapidly advancing synthetic biology.
In an accompanying commentary published in Science, Dr Thomas Inglesby and Dr Moritz Hanke of the Johns Hopkins Center for Health Security warned that the findings raise “urgent biosafety and biosecurity questions.”
The experts argued that the scientific community must now determine how AI-powered genome design can continue advancing without increasing the risk of accidental or deliberate misuse.
“It is no longer a question of whether generative viral genome design will exist,” they wrote, “but whether it can be developed without enabling serious harm.”
They cautioned that research involving viruses capable of causing disease in humans should not be pursued without robust safeguards.
Researchers stress safeguards
The Stanford team said multiple precautions were built into the project.
The AI systems were deliberately trained only on viruses that infect bacteria rather than humans or animals, reducing the possibility of generating harmful pathogens.
All laboratory work was conducted under controlled biosafety conditions.
Hie said existing scientific oversight frameworks already provide significant protection against misuse and argued that the technology has enormous potential to improve human health.
Beyond viruses
Although the study represents a major advance in synthetic biology, researchers stressed that designing viruses is far less complex than creating living organisms.
Viruses are not considered living entities because they cannot reproduce independently and require host cells to replicate.
The bacteriophage genomes designed during the study contain roughly 5,400 DNA base pairs, whereas the simplest known living cells require approximately 500,000 base pairs. By comparison, the human genome contains about three billion base pairs.
Even so, scientists believe increasingly capable AI models could eventually assist in designing far more complex biological systems.
Professor Marc Güell, a synthetic biology researcher at Pompeu Fabra University in Spain, described the study as a landmark moment.
“For the first time in history, we are beginning to design biology on a computer,” he said.
Also Read: Scientists create human embryos from skin DNA for the first time
He said the technology could eventually help develop more effective phage therapies, enzymes to treat inherited diseases and improved antibodies for cancer immunotherapy.
Professor Patrick Cai, Chair of Synthetic Genomics at the Manchester Institute of Biotechnology, said the findings suggest AI is beginning to understand biological design principles shaped by evolution.
“The significance extends far beyond phages,” he said. “It opens the door to AI-assisted genome writing.”
While researchers say practical applications remain years away, the study marks one of the clearest demonstrations yet that artificial intelligence is moving beyond analysing biology to actively designing it—an advance that could reshape medicine while intensifying debates over how powerful biotechnology should be governed.
