Scientists have used AI to design 16 new bacteriophages capable of targeting antibiotic-resistant E. coli, marking a breakthrough in biotechnology while sparking biosecurity debates. The research, published in Science, highlights both the potential for medical innovation and the risks of AI-driven pathogen design.
A study published in Science by a team of researchers trained an AI model named Evo on vast DNA datasets to generate synthetic bacteriophages—viruses that infect bacteria rather than humans. Out of 700,000 designs, 285 were tested in the lab, with 16 successfully halting E. coli growth, including antibiotic-resistant strains. The work, led by scientists at Stanford University, demonstrates how artificial intelligence could accelerate the development of targeted therapies against drug-resistant infections, a global health crisis linked to over 4.7 million deaths in 2021, according to the World Health Organization.
Scientists use AI to invent viruses never seen in
The AI-created phages represent a significant leap in synthetic biology, offering a potential solution to the growing threat of antibiotic resistance. However, the technology’s dual-use potential—its ability to design both therapeutic and harmful pathogens—has raised alarms among biosecurity experts. Isaac Bogoch, an infectious diseases specialist at the University of Toronto, noted that A.I.-designed viruses could have some potential benefits, such as the creation of targeted bacteriophages that could possibly help us tackle antibiotic-resistant infections in new ways,
as reported by Smithsonian Magazine.
To mitigate risks, the researchers excluded human pathogen data from the AI’s training set and implemented strict biosafety protocols. They also emphasized the complexity of designing viruses that could infect humans, noting that the AI-generated phages have the smallest and easiest genome to make, according to Tom Ellis, a professor of synthetic genome engineering at Imperial College London. The phage used in this research has an extremely small genome, meaning more complex viruses are much harder to design,
Ellis told Al Jazeera.
AI Model Trained In DNA Invents 16 New Viruses
Despite these precautions, critics argue that the technology’s accessibility could lower barriers for malicious actors. Dr. Moritz Hanke of the Johns Hopkins Center for Health Security warned that You could say, ‘Hey, genomic language model, make me an influenza genome that is modified to be more transmissible or to be more lethal,’
Hanke told the New York Times. Biosecurity leaders Dr. Thomas Inglesby and Dr. Moritz Hanke called for strict laws regarding future research and said it should be illegal to use similar generative techniques on pathogens affecting humans, animals, or agricultural crops, as outlined in an accompanying commentary in Science.
The study has ignited a debate about AI’s role in biological research. While some, like Patrick Cai of the University of Manchester, call it an important milestone, others warn of unchecked progress. The ability to custom-design new, functional viruses could lead to breakthroughs in treating superbugs, but it also underscores the urgent need for regulatory frameworks. The paper’s authors advocate for oversight, including restrictions on AI-generated pathogen research, as reported by Smithsonian Magazine.
Viruses on the Brain
The research underscores a broader tension in modern science: the faster we advance, the more we must grapple with the consequences. As AI’s capabilities expand, so too must our safeguards. The study’s authors stress that collaboration between scientists, policymakers, and the public is essential to ensure AI tools are used responsibly. We’d rather have a seat at the table than on the sidelines, said a researcher involved in the project, as reported by Smithsonian Magazine. For now, the creation of AI-designed viruses marks both a scientific triumph and a cautionary tale about the power—and peril—of emerging technologies.
