Artificial Intelligence Designs New Viruses... Lab-Reproducible
BBC – Researchers in the United States have announced success in using artificial intelligence technologies to design entirely new synthetic viruses, which were able to function and replicate in the laboratory. This achievement marks the first time that AI has been used to design the complete genome of a virus.
The researchers clarified that the 16 viruses developed target bacteria only and pose no risk to human health.
Several scientists described the achievement as a major turning point in biological research, potentially paving the way for new applications in disease treatment.
Brian Hie, an assistant professor at Stanford University, stated that the study represents “a new step in the level of complexity that generative AI can handle,” noting that this is the first time such AI has been used to design a complete genome capable of replication and performing functions within cells—a feat he described as a “new scientific frontier” for researchers.
However, the researchers applied this mechanism to what is known as the “language of life.” The AI models, named Evo 1 and Evo 2, were trained to analyze genetic codes rather than words, with the aim of predicting and designing genetic sequences.
Training these models relied on massive databases containing the genetic codes of viruses, bacteria, plants, and humans, before being fine-tuned to design a type of virus known as bacteriophages—viruses that infect specific types of bacteria and no others.
The Stanford research team selected 302 designs that the AI deemed most promising, and successfully synthesized them in the laboratory. Experiments showed that 16 of them were capable of eliminating Escherichia coli (E. coli) bacteria.
Samuel King, a PhD student who co-authored the study, said the team realized the experiment had succeeded in the early morning hours, when clear signs emerged that the new viruses were functioning naturally.
He explained that the researchers placed these bacteriophages on laboratory petri dishes containing a layer of bacteria, then monitored whether they would begin attacking them.
For his part, Brian said that once the team members saw the results, “the room where they were erupted spontaneously into applause,” highlighting the significance of the achievement.
Researchers believe that developing new bacteriophages could open the door to innovative treatments to combat bacterial infections that have become resistant to antibiotics—a major challenge facing modern medicine.
However, this achievement is not limited to the development of new viruses; it also points to the growing ability of AI to design biological systems that go beyond what exists in nature, a field known as synthetic biology.
Brian believes this technology could bring about a qualitative leap in healthcare by accelerating the development of innovative drugs and treatments that could significantly improve human health on a broad scale.
This progress has raised growing concerns about the potential misuse of the technology to design new pathogens for harmful purposes.
In a commentary published alongside the study in the journal Science, Dr. Thomas Inglesby and Dr. Moritz Hanke of the Johns Hopkins Center for Health Security stated that the study’s findings raise “urgent questions regarding safety and biosecurity.”
The two researchers noted that the question is no longer whether AI technologies can design viral genomes, but rather how to ensure their use without causing “serious harm.”
The researchers, for example, emphasized that viruses with the potential to cause disease should not be allowed to be developed.
