Published: 21:26 BST, 6 August 2026 | Updated: 22:16 BST, 6 August 2026<\/p>\n
Scientists have used\u00a0Artificial Intelligence\u00a0to design new viruses capable of killing cells in the laboratory.<\/p>\n
It marks the first time that the new technology has been used to successfully generate whole genomes, or the full set of genetic instructions needed to build a working organism.\u00a0<\/p>\n
Supporters of the work said it offered hope for developing new treatments, but critics quickly warned it raised 'urgent' safety and security concerns.\u00a0\u00a0<\/p>\n
In their study, researchers at Stanford University, California, used the technology to create a genome for a virus that infects bacteria.<\/p>\n
The AI suggested thousands of genomes, and the researchers created 302 in the lab before exposing them to bacteria.<\/p>\n
Overall, 16 of the viruses suggested by the AI were able to kill E.coli.<\/p>\n
The viruses they created were bacteriophages, which only infect bacteria and are not able to infect human, animal or plant cells.\u00a0<\/p>\n
Revealing the results, Dr Brian Hie, a chemical engineer behind the research, said: 'In this case, we wanted the model to generate the entire genome end-to-end in a single left-to-right pass. We didn't add anything.'<\/p>\n
Scientists have used AI to design a new virus that was capable of infecting other cells\u00a0<\/p>\n
The research was published alongside an accompanying article warning over the potential risks of the advance.\u00a0<\/p>\n
Written by Johns Hopkins experts Dr Thomas Inglesby and Dr Maurice Hanke, it warned:\u00a0'Although this is promising for life sciences applications, it also raises urgent biosafety and biosecurity questions.<\/p>\n
'The ability to compose viral genomes using generative AI now exists; the governance to safely steer it does not.'<\/p>\n
For the research, published in Science, scientists used the AI tools Evo1 and Evo2.\u00a0<\/p>\n
These work in a similar way to large language model chatbots ChatGPT and Grok, except they have been trained on genetic codes rather than written text.<\/p>\n