AI-Designed Viruses Target Bacteria

Stanford ‍researchers have ⁢developed an AI model named⁣ Evo that can generate DNA sequences ⁤for viruses that infect bacteria.Evo was trained ⁤using millions ‍of viral genome sequences. The study, available as a
⁢
preprint, is currently awaiting peer review.

⁣ ‍ The researchers focused on creating variants of the virus Phix174, known for infecting strains of the E. coli bacterium. The team synthesized 302 new viruses based on Phix174.
⁣

Of these, 16 were⁢ viable and capable of infecting and killing E. coli. Alarmingly, some of the AI-generated viruses proved more lethal than the naturally occurring Phix174.

Implications for Bioweapon Development

⁣ The experiment raises concerns about the potential for AI to be used in the creation of biological weapons. The researchers⁣ warn that if AI accelerates the design of biological weapons, the United States⁢ and ⁤other nations ⁣must accelerate their response capabilities.

“We cannot stop new threats⁤ generated by AI. The real challenge is to stay ⁤ahead of them,” the Feldmans warn, emphasizing the need to leverage AI for defensive purposes, such as designing antibodies, antiviral agents, and vaccines.
‍

Using AI for Defense

‍ The development of AI-generated viruses underscores the⁤ dual-use nature of ⁢AI in biotechnology.While AI can be used to create harmful ‍pathogens, it can also be a powerful‍ tool for developing countermeasures.

Researchers and governments are exploring the use of AI to ⁣rapidly identify potential threats,design effective treatments,and develop preventative measures against infectious diseases.