A startup company called Radical Numerics says it has used an AI-generated virus for the first time to treat a bacterial infection in a human. The claim, however, is being met with some skepticism due to a lack of detail and peer review.
While Radical Numerics has posted a technical report on its website, the company has not revealed any information about the patient or the doctors running the treatment, citing confidentiality agreements. The U.S. Food and Drug Administration authorized the treatment under a compassionate-use arrangement, according to company officials.
CEO Eric Nguyen told the New York Times the company takes patient confidentiality and maintaining the trust of its clinical partners very seriously. He added: “We intend to publish our scientific findings through peer-reviewed research, with any future disclosure of clinical details subject to appropriate consent.”
Radical Numerics researchers said an AI-generated virus, called a phage, was administered to a patient suffering from a serious bacterial infection. Phages attack bacteria and can kill off the microbes causing the infection. The company has not disclosed the patient’s treatment outcome.
If Radical Numerics’ use of AI proves to be successful, it could pave the way for individualized treatment of bacterial infections. Antibiotic resistance is becoming an increasingly serious problem. Clinics that specialize in phage therapy are opening up to help patients deal with infections that resist multiple drugs.
The Radical Numerics announcement comes as the use of AI in biology comes under increasing scrutiny. In August, scientists at Stanford University and the Arc Institute reported using AI to recognize DNA patterns in nature and then use that information to write new recipes for entirely new viruses. The scientists used an AI model called Evo that works similarly to existing AI models but concentrates on genetic sequences instead of words. The researchers tested 285 AI-generated viral genomes, but only 16 produced viable phages. Still, the creation of viruses not found in nature is giving many pause.
While the viruses developed with Evo were bacteriophages that infect bacteria rather than humans, the development prompted fears that AI could be used to create new biological weapons ranging from deadly poisons to unstoppable pandemics.
Dr. Jeff Coller, director of the RNA Innovation Center at Johns Hopkins University, is among those alarmed. In an October 12 New York Times op-ed, Coller warned: “We’re entering an era in which a machine can design a pathogen faster than we can create the countermeasure to fight it.”
Coller says the government needs to create the permanent equivalent of Operation Warp Speed, created to fight the COVID-19 pandemic, in order to develop countermeasures against AI-created pathogens quickly and ensure biosecurity.
