Scientists have used artificial intelligence to design entire viral genomes and synthesize bacteriophages that can infect and kill antibiotic-resistant E. coli, in a preprint that underscores AI’s accelerating reach into synthetic biology. Trained on millions of phage sequences and steered toward a ΦX174-like template, the Evo 1 and Evo 2 models generated hundreds of candidates; 16 showed host specificity, and combinations cleared three E. coli strains that resisted wild-type ΦX174. The results mark a milestone—AI writing coherent, genome-scale code—with potential to augment phage therapy against drug-resistant infections. But the work, not yet peer reviewed, raises biosafety and regulatory questions as policymakers weigh controls on frontier bio-AI systems. Commercial prospects hinge on reproducibility, safety, and a path through evolving oversight frameworks for engineered pathogens.
Related articles:
Engineered bacteriophages for treatment of a patient with a disseminated drug-resistant Mycobacterium abscessus infection
Highly accurate protein structure prediction with AlphaFold
Highly accurate protein structure prediction for the human proteome
The Bletchley Declaration by Countries Attending the AI Safety Summit, 1–2 November 2023
Biodefense in the Age of Synthetic Biology
