US researchers have successfully used AI to design 16 novel, fully functional viruses that infect bacteria, marking a milestone in synthetic biology.

Key facts
- •The 16 AI-designed bacteriophages were tested on E. coli bacteria.
- •The AI models were trained on genetic sequences from viruses, bacteria, plants, and humans.
- •The phage genomes designed by the AI are approximately 5,400 base pairs long.
- •Researchers at Stanford University led the study, which was published in the journal Science.
- •Experts warn that the same technology could potentially be misused to create harmful diseases.
Stanford University researchers have used generative AI to design 16 novel viruses capable of replicating in a laboratory setting. These viruses, known as bacteriophages, were engineered to infect specific bacteria and pose no threat to humans. This marks the first time that whole genomes have been successfully designed by AI, a development described by experts as a significant turning point in science.
By the numbers
How the AI Designed the Viruses
The research team utilized AI models named Evo1 and Evo2, which function similarly to large language models like ChatGPT. Instead of predicting text, these models were trained on genetic codes from viruses, bacteria, plants, and humans to predict the 'language of life.' The researchers synthesized 302 of the AI-generated designs, 16 of which proved effective at killing E. coli bacteria in laboratory tests.
Safety and Security Concerns
While the technology could lead to new treatments for antibiotic-resistant infections, it has also prompted warnings regarding biosafety and biosecurity. Dr. Thomas Inglesby and Dr. Moritz Hanke of the Johns Hopkins Center for Health Security noted that the findings raise urgent questions about the potential for malicious use. In response, the Stanford team stated they took safety precautions by excluding viruses that infect complex organisms from their training data and conducting the research in a secure laboratory.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by BBC Health.


