Stanford's AI creates 16 new viruses to fight superbugs

Stanford University and the Arc Institute have achieved a milestone in synthetic biology by using AI to design 16 entirely new viruses capable of infecting bacteria, including antibiotic-resistant strains. The breakthrough leveraged two foundational AI models—Evo 1 and Evo 2—trained on millions of genomes across all domains of life to identify evolutionary patterns and generate novel viral architectures.

The researchers focused on bacteriophages (viruses that infect bacteria), using the Phi X-174 bacteriophage as a functional template. Their AI models generated thousands of new genomes compatible with infecting *E. coli*, though only 16 of 300 synthesized genomes produced functional viruses with entirely novel sequences and gene organizations. Critically, these AI-designed viruses successfully overcame bacterial resistance, demonstrating potential for next-generation phage therapies against rapidly evolving pathogens. Published in *Science*, the study underscores both the promise of AI-driven synthetic biology and the dual-use risks of such technologies.