AI-Designed Novel Viruses Spark Urgent Biosecurity Debate

Researchers have used artificial intelligence to design and create novel viruses entirely new to science. These viruses are bacteriophages, which are more than 90 percent similar to existing bacteria-infecting viruses and only pose a danger to Escherichia coli. While the breakthrough offers potential for gene therapies and treating antibiotic-resistant bacteria, it has triggered an urgent debate over biosecurity.

Sentinel Bio Inc. and other experts warn that AI-enabled biotechnology is advancing faster than regulation, potentially allowing bad actors to design novel bioweapons or increase viral virulence. Microsoft recently discovered AI-generated toxic protein sequences that bypassed safety screenings, necessitating software patches to close the gap.

To mitigate these risks, some experts propose a tiered access system for AI training data and more rigorous screening of synthetic DNA and RNA sequences. However, policymakers in the United States have not yet implemented formal safeguards. This leaves the responsible release of powerful biological models largely to the goodwill of the scientific community.