In a significant milestone for synthetic biology, researchers in the United States have successfully utilized artificial intelligence to design brand new, fully functional viruses. This achievement marks the first time that entire genomes have been successfully designed by AI, resulting in 16 novel viruses that are capable of replicating in a laboratory environment. These specific viruses were engineered to infect bacteria and pose no threat to human beings.
While AI tools have previously been employed to design new antibiotics, this project represents a substantial increase in complexity. Brian Hie, an assistant professor at Stanford University, noted that this is the first instance where generative AI has been used to design a complete, viable genome that can replicate and perform functions within cells. Hie described the work as "new territory" for the research team, representing a significant step forward in what is designable by generative AI.
The technology behind this breakthrough operates similarly to large language models like ChatGPT, which predict sequences of text. In this case, the AI models—named Evo1 and Evo2—were trained on genetic codes derived from viruses, bacteria, plants, and humans to predict the "language of life" rather than words. The models were specifically refined to produce bacteriophages, which are viruses that target and infect only specific species of bacteria.
To test the viability of these digital designs, the Stanford researchers selected the 302 most promising AI-generated candidates and synthesized them in the laboratory. Of these, 16 proved effective at killing E. coli bacteria. Samuel King, a PhD student in the lab, recounted the moment the team realized the phages were working in the early hours of the morning. As they observed clear spots on petri dishes where the viruses were consuming the bacteria, the team felt immense excitement, and Hie recalled that the room "spontaneously burst into applause" when the results were shared.
The development of new phage therapies is viewed as a potential solution to the growing crisis of bacterial infections that are resistant to traditional antibiotics. Beyond this immediate application, the breakthrough demonstrates the capacity of AI to design new biology that extends beyond the natural world. Prof. Marc Güell, from the synthetic biology lab at Pompeu Fabra University in Spain, called the study a "very significant turning point," noting that for the first time in history, scientists are beginning to design biology on a computer.
Prof. Patrick Cai, chair of synthetic genomics at the Manchester Institute of Biotechnology, added that the study is an "important milestone" suggesting that genome language models are learning design principles encoded by evolution.
Despite the potential to "massively improve human health," the research has prompted concerns regarding the potential for malicious use. In a commentary published in the journal Science, Dr. Thomas Inglesby and Dr. Moritz Hanke from the Center for Health Security at Johns Hopkins University warned that the findings raise "urgent biosafety and biosecurity questions." They argued that the focus must now be on whether this technology can be used without "enabling serious harm," specifically stating that the creation of new viruses with the potential to cause disease "should not be pursued."
The researchers emphasized that they took significant steps to maximize safety, including performing the work in a secure laboratory, focusing on phages rather than human-infecting viruses, and excluding viruses that could infect complex organisms from their training database. Hie noted that while generating living organisms would require a massive leap in complexity—as the phage genome is only 5,400 base pairs long compared to the 500,000 base pairs of the smallest living cell—the team is "definitely interested in working towards" that goal in the future.
AI designs antibiotics for gonorrhoea and MRSA superbugs
