Scientists Used AI to Create 16 New Viruses
| Source: Wired AI
Tags: Evo 2, Arc Institute, Stanford, synthetic biology, bacteriophage, biosecurity, antibiotic resistance, computational biology
Stanford and Arc Institute researchers used Evo 1 and Evo 2 AI models to design 16 previously unknown bacteriophages capable of infecting E. coli — the first time AI created entirely novel viral genomes from evolutionary patterns, with antibiotic-alternative potential and explicit biosecurity implications.
Details
Stanford University and Arc Institute researchers used Evo 1 and Evo 2 — foundation AI models trained on millions of genomes across all biological domains — to design bacteriophages entirely from scratch. Rather than modifying known viral sequences, the AI generated novel DNA architectures by learning evolutionary patterns: which gene organizations are functional, which sequences are conserved, and what biological constraints allow an organism to survive. Using bacteriophage phiX174 (which infects E. coli) as a guide — not a template — the AI generated thousands of candidate genomes. A computational screening framework cut the pool to 300 for chemical synthesis. Of these, 16 showed strong E. coli-killing activity in lab tests, with some outperforming the natural reference phage itself. The medical significance is substantial: bacteriophages are a promising alternative to antibiotics for drug-resistant infections. AI-accelerated genome design could dramatically speed the development of targeted phage therapies for pathogens like MRSA and resistant E. coli strains. The biosecurity concern is explicit: scientists have synthesized known viruses for years, but AI generating novel functional genomes from learned evolutionary patterns alone represents a qualitatively different capability. Current regulatory frameworks govern known pathogen synthesis and do not adequately address AI-designed novel organisms.