AI Designs 16 New Functional Viruses in Lab First
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- Stanford University researchers used generative AI models Evo1 and Evo2 — trained on genetic codes from viruses, bacteria, plants, and people — to design 16 novel bacteriophages that successfully killed E. coli, the first whole genomes designed by AI.
- Of 302 AI-designed candidates synthesized in the lab, 16 proved effective, with researcher Brian Hie calling it a leap in 'complexity that's designable by generative AI.'
- Dr. Thomas Inglesby and Dr. Moritz Hanke of Johns Hopkins's Center for Health Security warned the findings raise 'urgent biosafety and biosecurity questions,' arguing viruses with disease potential 'should not be pursued.'
- Prof Marc Güell of Pompeu Fabra University called the study a 'very significant turning point' — 'the first time in history, we are beginning to design biology on a computer.'
- The bacteriophage genome is roughly 5,400 base pairs long, compared with 500,000 for the smallest living cell and 3 billion for humans; Hie said designing simple organisms is 'a lot of work, but not impossible.'
- Researchers say the breakthrough could enable new phage therapies for antibiotic-resistant infections, while Hie argues the same tools could 'massively improve human health' through new drugs and therapies.
Why it matters: The 16 successful bacteriophages mark the first AI-written genomes to function in a lab, a stepping stone toward AI-designed living organisms (the smallest cell genome is roughly 100× larger than the phage's). The same capability that promises phage therapies for antibiotic-resistant infections also gives regulators a concrete example — not a thought experiment — of why biosecurity guardrails must catch up before the technology scales.
