AI Chatbot Aids Investigation in Illinois Salmonella Outbreak
Health officials in Illinois faced a challenging outbreak linked to a county fair and turned to an AI chatbot for assistance. The outbreak first came to light when the Brown County sheriff reported an unusual number of potential jurors with stomach issues on August 5, 2024. A week later, the state health department confirmed a case of Salmonella enterica serotype Agbeni, prompting a broader investigation.
Officials identified 13 cases of the infection, with seven confirmed and six probable cases. The common factor among all affected individuals was attendance at the Brown County fair, which attracts around 36,000 visitors annually. The fair had concluded by the time the connection was made, complicating efforts to trace the source of the outbreak.
Initial hypotheses focused on food vendors, as Salmonella is typically associated with foodborne illnesses. However, four of the affected individuals had not consumed any food at the fair, leading investigators to consider other possibilities. Notably, all those infected had consumed canned beer from the fair's single beer tent. The beer was stored in a makeshift cooler made from non-food-grade materials, which was inadequately cleaned and maintained throughout the event.
Health officials hypothesized that the cooler could have been contaminated with Salmonella, potentially spreading the bacteria to the beer cans. However, with the cooler no longer available for testing, this remained speculative. To bolster their hypothesis, investigators consulted ChatGPT, an AI chatbot, for insights. They queried the chatbot on the likelihood of Salmonella growth in such conditions and sought comparisons with documented outbreaks.
While some answers could have been found through traditional research methods, the chatbot confirmed the cooler as a plausible source of contamination. This led to the implementation of new sanitation protocols for coolers at future events. Katherine Houser, the county health official who authored the report, noted that AI provided rapid situational awareness, though she acknowledged the limitations and potential inaccuracies of AI-generated information.
The case highlights both the potential and the challenges of using AI in public health investigations. While the chatbot's input was deemed helpful, the necessity of critically reviewing AI-generated content remains crucial. Ultimately, the exact cause of the outbreak remains uncertain, but the introduction of improved sanitation measures is a positive outcome.