What if scientists could anticipate the likelihood of an outbreak before it starts? What if, instead of scrambling to contain an emerging threat, we could prevent it?
In the last year, waves of illnesses such as Marburg virus, Mpox and avian influenza (H5N1) have underscored the need for enhanced preparedness to prevent future outbreaks. While the world continues to grapple with the long-term effects of COVID-19, scientists are working to harness the full potential of artificial intelligence in global health security through AI-driven modeling and predictive analytics — hoping to foresee potential outbreaks, assess risk factors and create early warning systems that can save lives.
Researchers at the University of Florida’s Emerging Pathogens Institute, for example, published an algorithm capable of predicting which COVID-19 variant in circulation today is most likely to become dominant in the next three months. This algorithm was used to spot new variants of concern, and it correctly identified 11 out of 11 variants a full 10 weeks before they were officially labeled by the Centers for Disease Control and Prevention.
By training these algorithms on publicly available genetic sequences of SARS-CoV-2, scientists can continue to predict which mutations will pose the greatest threat to public health.
The potential of AI in the lab stretches far beyond COVID-19, however. When applied to other infectious diseases like HIV/AIDS, AI-driven pattern recognition can uncover risk factors that contribute to differences in treatments and outcomes, and even locate where transmission remains stubbornly high.
Months prior to the COVID-19 pandemic, global change studies demonstrated AI’s potential in informing viral emergence trends to potentially beat the next outbreak. Through built datasets and experiential learning, the application of AI for cross-species transmission has been used to understand the transmission of zoonotic diseases, primarily vector-borne diseases, as well as the impacts of global change on disease emergence and spread.
Infectious disease research can leverage both novel and existing data across systems to answer computational questions about the fundamental issues of anticipating emerging pathogens. Such studies are multidisciplinary, involving groups of scientists learning more about host-virus networks, as well as data fluency and knowledge translation.
Given the vast amounts of data being processed today, there is a critical need to enhance the quality of the technology used to analyze it, and to communicate this need to the broader scientific community. Integrating AI into the study of infectious diseases emphasizes the need to strengthen research capabilities for the future, which means pushing the envelope on scientific advancement as it relates to all types of pathogens.
Bacterial diseases like anthrax, for instance, also benefit from AI-enhanced detection techniques. In Southwest Texas, where anthrax transmission is intertwined with how animals interact with their landscape, machine learning is making a world of a difference. Researchers are developing an AI-based model that can forecast the likelihood of an outbreak based on what the first few months of animal cases look like. The goal? To develop a way to determine the likelihood of an outbreak with two months’ notice.
Florida is a top-ranking state for U.S. crop production and serves as a model for disease management in parts of the world with similar climates. Global plant health, too, can benefit from the smart solutions offered by AI. Farmers can use AI to design and build smart agricultural systems tailored for disease management and minimize the impact of disasters on agricultural systems. Agricultural workers can use AI models to support food security and maximize the way growers sanitize tools and practice crop rotation.
By feeding AI models decades of data, algorithms become familiar with varying levels of greenness that correspond to differing amounts of vegetation in an area. After showing the algorithm partial patterns, researchers teach it to find matches from previous years and later fill in the gaps of incomplete data to inform predictions for future years.
These applications are just the beginning. AI’s potential to detect and prevent disease is limitless, and allows us to address public health in a manner that is proactive, not just reactive. The fight to prevent the next pandemic can be different.
Marco Salemi, Ph.D., is the interim director of the University of Florida Emerging Pathogens Institute and a professor in the University of Florida College of Medicine’s Department of Pathology, Immunology and Laboratory Medicine.
