Play store
Scientists in the United States have used artificial intelligence (AI) to design fully functional viruses from scratch for the first time, marking a major breakthrough in synthetic biology while raising fresh concerns over biosafety and biosecurity.
The research, led by scientists at Stanford University and published in the journal Science, successfully created 16 new bacteriophages, or phages, viruses that infect bacteria but are harmless to humans. The AI-designed viruses were shown to replicate in laboratory conditions and effectively kill E. coli bacteria.
Researchers said the achievement represents the first successful use of generative AI to design complete viral genomes capable of functioning inside living cells.
“This is the first time generative AI has been used to design a complete genome that can replicate and perform biological functions,” said Brian Hie, assistant professor at Stanford University. “It is a new frontier for AI-driven biology.”
The study used AI models known as Evo1 and Evo2, which were trained on genetic sequences from viruses, bacteria, plants and humans. Similar to large language models that predict words in a sentence, these systems learn the “language of life” by predicting genetic sequences.
Scientists generated 302 AI-designed viral genomes and synthesized them in the laboratory. Sixteen of those proved capable of infecting and destroying E. coli bacteria.
Researchers believe the technology could accelerate the development of phage therapies to combat antibiotic-resistant infections, offering a promising alternative as drug resistance continues to grow worldwide.
The breakthrough also signals a new era in synthetic biology, where AI may eventually help design novel medicines, therapeutic enzymes, antibodies and other biological tools. However, experts have warned that the same technology could be misused to create harmful pathogens.
In an accompanying commentary in Science, Dr. Thomas Inglesby and Dr. Moritz Hanke of the Johns Hopkins Center for Health Security said the research raises “urgent biosafety and biosecurity questions.” They stressed that the challenge is no longer whether AI can design viral genomes, but how such technology can be governed to prevent misuse.
To reduce risks, the Stanford team excluded viruses capable of infecting humans and other complex organisms from the AI’s training data. The research was conducted exclusively on bacteriophages under secure laboratory conditions.
Although researchers say designing living organisms remains far more complex, they acknowledge that AI’s ability to write biological genomes is advancing rapidly.
Experts have described the work as a landmark in computational biology.
Marc Güell, professor of synthetic biology at Pompeu Fabra University in Spain, called it “a very significant turning point,” saying it marks the first time scientists have begun designing biology entirely on computers.
Patrick Cai, chair of synthetic genomics at the Manchester Institute of Biotechnology, said the findings suggest AI is beginning to understand the genetic design principles shaped by evolution, opening the door to AI-assisted genome engineering.
The study highlights both the enormous medical potential of AI-designed biology and the need for robust international oversight to ensure the technology is developed safely and responsibly.