AI Creates 16 New Functional Viruses Targeting Bacteria

For the first time, an AI system has designed a series of entirely new, functional viruses. Researchers at Stanford University and the Arc Institute used two foundational computational biology models — Evo 1 and Evo 2 — trained on millions of genomes spanning all domains of life, to generate novel bacteriophage genomes. The study used the bacteriophage Phi X-174, which infects Escherichia coli, as a reference architecture — not as a template to copy, but as a guide for the AI to produce genomes with compatible functional organization. The algorithms generated thousands of candidate genomes with gene layouts and regulatory structures consistent with infecting E. coli, while using DNA sequences that differed substantially from any known natural bacteriophage. Researchers narrowed this pool down to 300 candidates based on biological criteria, synthesized each molecule by molecule in the lab, and introduced them into E. coli. Of the 300, 16 produced fully functional bacteriophages. The breakthrough holds significant potential for addressing antibiotic-resistant bacterial infections, as bacteriophages — which infect only bacteria — are considered a promising alternative to antibiotics. However, the research also raises regulatory concerns, as the technology's pace is outstripping existing biosecurity frameworks and could theoretically be misused to design biological weapons.