A Stanford-led research team reported using generative AI to design functional viruses that infect E. coli, creating genomes not previously seen in nature. The work was published in Science alongside a biosecurity commentary. The research used genome language models to design complete viral genomes. The resulting bacteriophages infected and killed E. coli.
The team described the work as the first test of whether such models could generate entire functional genomes. The researchers reported safety precautions in the study. The source record attributes statements, allegations and official descriptions to the people or institutions making them.
Bacteriophages infect bacteria rather than people. Synthetic biology already uses screening and laboratory safeguards. AI-generated sequences can challenge oversight systems written around known organisms or conventional modification. This background supplies chronology and operating context without deciding the outcome still under review.
A companion Science commentary said existing governance was insufficient for generative genomics. The current U.S. high-risk research policy focuses on federally funded work and harmful biological agents. Counts, dates and legal status remain tied to the checked source record because later reporting can revise preliminary information.
The evidentiary limit is specific: The accessible reporting did not establish performance beyond the reported laboratory setting or show that a human pathogen had been created. The available reporting does not support an inference beyond that boundary.
The next record points are independent replication and methodological review and specific biosafety and sequence-screening rules for generative genome tools. Each can be checked against later official documents or independent reporting.
The verified chronology is therefore limited to the following: The research used genome language models to design complete viral genomes. The resulting bacteriophages infected and killed E. coli. The team described the work as the first test of whether such models could generate entire functional genomes. The researchers reported safety precautions in the study. A companion Science commentary said existing governance was insufficient for generative genomics. The current U.S. high-risk research policy focuses on federally funded work and harmful biological agents. The relevant context is Bacteriophages infect bacteria rather than people. Synthetic biology already uses screening and laboratory safeguards. AI-generated sequences can challenge oversight systems written around known organisms or conventional modification. These details state what the sources establish and preserve what remains unknown.