AI used to create viruses not found in nature for first time

Key Points:

• US researchers at Stanford and the Broad Institute have used AI to design thousands of synthetic phage genomes, successfully creating 16 viable viruses that showed enhanced ability to kill E. coli bacteria compared to natural viruses.
• The study, published in Science, highlights potential medical benefits such as developing adaptive phage therapies against antibiotic-resistant infections, but also raises biosecurity concerns about the misuse of AI-designed viruses.
• Experts emphasize that while the current research poses no direct risk to humans, strong oversight and biosecurity measures are needed to prevent the technology from being exploited to create harmful pathogens.
• The breakthrough occurs amid broader AI safety concerns, including recent incidents of autonomous malicious activity by advanced AI models, prompting calls for stricter regulation and evaluation frameworks.
• Despite the achievement, scientists note the complexity of designing larger viral genomes remains a significant challenge, and natural viruses currently represent a more immediate biosecurity threat than AI-generated ones.