
Stanford University and Arc Institute researchers used AI models Evo 1 and Evo 2 to design complete genomes of bacteriophages. Of 285 AI-generated designs physically synthesised and tested, 16 produced functioning viruses.…
Stanford University and Arc Institute researchers used AI models Evo 1 and Evo 2 to design complete genomes of bacteriophages. Of 285 AI-generated designs physically synthesised and tested, 16 produced functioning viruses. Some AI-designed phages overcame bacterial resistance that defeated the original virus. The Hindu reports that while humans have synthesised and modified viral genomes for decades, this is the first time AI has entered the design stage, proposing entire genome sequences rather than individual mutations.
The advance has clear medical promise for phage therapy against antibiotic-resistant bacteria. AI-designed phages successfully attacked E. coli strains where the original virus failed. However, the same technology that can design therapeutic viruses could accelerate the creation of dangerous pathogens. The finding that AI can identify multiple genetic changes working together across a genome highlights both opportunities and biosecurity dilemmas that remain unresolved.
Headlines screaming 'AI creates viruses' miss the nuance. AI designed a genome for a harmless bacteria-infecting virus, not a human pathogen. Yet the hand-waving dismissals are equally lazy. The same generative biology that helps design phages for antibiotic resistance can, tomorrow, propose sequences for more dangerous viruses. The real test is whether Evo 2 could design a coronavirus spike protein that evades current immunity. That experiment is already possible. Who runs it first, and under what oversight, will settle the argument.
Source: thehindu.com
This story was synthesised by AI from the source linked above.