A recent study conducted by the University of Maryland School of Medicine investigated their new artificial intelligence (AI) device’s ability to detect bird flu risk. The favorable findings are reported in a University of Maryland School of Medicinerelease by Deborah Kotz.
The scientists tested the AI on over 13,000 records from emergency visits in the University’s hospitals. Patients were from different areas: urban, suburban, and rural. Each patient was seen for respiratory symptoms parallel to H5N1 bird flu infection symptoms. According to Kotz, “The goal was to assess how well generative AI could find high-risk patients who may have been overlooked at the time of initial treatment.”
The tool, ‘GPT-4 Turbo,’ proved to be successful at quickly and cheaply marking patients who had possible contact with the virus because their professions require them to work with animals. Researchers were then able to do a deeper analysis to verify which suspected patients actually were exposed to animals that could be carrying bird flu; however, they were not tested for the virus.
It is important to note that GPT-4 Turbo should not replace clinician assessment, it should only be used as a tool.
