Brian Hie, head of the lab at the Arc Institute, remarked on the notable experience of observing an AI-generated sphere. In total, 16 out of 302 AI-designed phage constructs were successful in replicating, leading to the elimination of bacterial hosts. J. Craig Venter, recognized for producing some of the first organisms with lab-created DNA nearly two decades ago, described the AI techniques as a more rapid iteration of traditional trial-and-error methods. Venter referenced his team’s 2008 achievement of creating a bacterium with a lab-printed genome, which involved extensive testing of various gene combinations.
The potential for AI to revolutionize biology is drawing significant interest, particularly due to advancements that have already earned a Nobel Prize in 2024 for accurately predicting protein structures. Investors are also heavily financing projects geared towards drug discovery, exemplified by Lila, a Boston-based company that recently secured $235 million to develop AI-operated automated laboratories.
The applications of computer-designed viruses may extend into commercial realms. Phage therapy, which utilizes viruses to treat serious bacterial infections, has piqued interest. There are also ongoing tests aimed at eradicating bacterial infections in crops, such as black rot in cabbage. Samuel King, a student who led the project in Hie’s lab, observed that gene therapy frequently employs viruses to deliver genes into patients. He posited that AI could lead to the development of more effective viral vectors.
Although the Stanford researchers have intentionally excluded human pathogens from their AI training, they acknowledged the potential risks. The technology could be misapplied by others—whether from curiosity, good intentions, or malicious aims—resulting in unforeseen consequences when applied to human pathogens.
Source: https://www.technologyreview.com/2025/09/17/1123801/ai-virus-bacteriophage-life/