Scientists have leveraged artificial intelligence (AI) to construct complete viral genomes in the lab, specifically targeting bacteriophages, which infect bacteria. A Microsoft-led study, spearheaded by Bruce J. Wittmann, highlighted concerns over AI-designed toxins that can bypass standard DNA safety checks. Utilizing advanced genome-language models trained on extensive data, researchers crafted hundreds of candidate phage genomes, successfully cultivating 16 viruses. This dual-use research poses significant security risks, as it can benefit medical advances like tailored phage therapies for antibiotic-resistant infections but also facilitate biological attacks. Safety measures include improving DNA screening processes and implementing federal policies linking funding to stringent nucleic-acid screening. The International Biosecurity and Biosafety Initiative aims to update screening norms as AI technology evolves. Experts advocate for biosecurity policies to mitigate misuse, emphasizing the need for oversight in this rapidly advancing field. The findings reveal both the potential and pitfalls of AI in genome design, underscoring the importance of responsible innovation.
Source link
Share
Read more