AI Designs Novel Viruses, Sparking Biosafety Concerns

Stanford University researchers have utilized artificial intelligence to design 16 novel viruses, a breakthrough that could offer new avenues for combating antibiotic-resistant infections. The AI models, Evo1 and Evo2, were trained to predict genetic codes and successfully generated hundreds of designs for bacteriophages capable of infecting E. coli bacteria, with 16 proving viable in the lab. This marks the first instance of AI designing complete viral genomes.

These newly designed bacteriophages hold the potential to serve as an alternative to traditional antibiotics, which are becoming less effective against increasingly resistant bacteria. Brian Hie, an assistant professor at Stanford University, described the development as a significant step in the complexity of designs achievable with generative AI, noting it was new territory to design a complete, self-replicating genome.

However, this advancement also raises significant biosafety and biosecurity concerns. Experts have warned that the technology's ability to create novel life forms, a concept within synthetic biology, could be exploited for malicious purposes, including the creation of bioweapons. In a commentary accompanying the study's publication in the journal Science, Dr. Thomas Inglesby and Dr. Moritz Hanke of the Johns Hopkins Center for Health Security emphasized the urgent nature of these risks. They stated that the question is no longer if generative viral genome design will exist, but rather how to prevent its use from causing serious harm, adding that viruses capable of causing disease should not be pursued.

Hie stressed that the research team took precautions by excluding viruses that could infect complex organisms from the AI's training data and focused solely on bacteriophages. He also confirmed that all experiments were conducted in a secure laboratory environment, asserting that safeguards are in place to ensure the technology is used for beneficial purposes.