In research that pushes the boundaries of synthetic biology, scientists have used artificial intelligence to create 16 completely new, viable viruses. These genetic codes, which differ significantly from all known natural viruses, represent a promising step toward developing new therapies to combat antibiotic-resistant bacteria, but at the same time raise serious concerns about potential misuse, according to CNN.
The Language of Life as Inspiration for AI
The study, published last week in the prestigious journal Science, details how researchers utilized an AI tool called Evo. Unlike large language models like those underpinning ChatGPT, which are trained on vast amounts of text, Evo built its "language base" on genetic sequences from millions of sources and, as stated in the study, "from all domains of life."
The team from Stanford University and the Arc Institute in the US then tasked the AI with generating thousands of genome combinations within strictly defined parameters, specifically those that could fit into an E. coli bacterial cell. The goal was for the AI to "learn the evolutionary constraints" of natural genomes and attempt to surpass them.
From Thousands of Designs to 16 Viable Viruses
Of the thousands of genetic blueprints produced by the AI, scientists synthesized and tested about 300 of them in the lab. The results were fascinating: 16 of these genomes resulted in viable viruses. All newly created viruses belong to the bacteriophage group, meaning they are specialized in infecting bacteria and cannot infect human cells.
What particularly surprised the researchers was that one of the new viruses possessed a trait that was "evolutionarily very distant" from anything known. This suggests that the AI managed to drive genetic changes that would perhaps take natural evolution millions of years to prove possible.
The Promise of Phage Therapy
The practical application of this discovery was immediately tested. A mixture of the newly created viruses managed to overcome bacterial resistance in some strains of E. coli, which had not been possible before. This is especially significant in the context of the growing global threat of antimicrobial resistance, where traditional antibiotics are becoming ineffective.
"This opens the way for the development of adaptable and resilient bacteriophage therapies against rapidly evolving pathogens," the study authors wrote in their paper.
Bacteriophage therapy, although nearly a century old, is gaining importance again precisely because of the antibiotic crisis. The ability to rapidly design customized viruses with the help of AI to combat specific bacterial infections could revolutionize medicine.
The Double-Edged Sword of Progress
Despite the medical potential, the study inevitably also raises questions about the dual-use nature of the technology. Although the created viruses are harmless to humans, the very fact that AI can be used to create entirely new forms of life, surpassing millions of years of evolution, sparks ethical and safety debates. The possibility that similar methods could be misused to create pathogens with undesirable traits presents a challenge for which an adequate regulatory framework does not yet exist.