Online restaurant reviews are being utilized to identify food poisoning outbreaks, as part of a recent study by the UK Health Security Agency (UKHSA). Using artificial intelligence (AI), researchers analyzed reviews from platforms like Yelp, focusing on comments that indicated gastrointestinal (GI) symptoms such as nausea, diarrhea, and vomiting.
The AI proved effective, successfully identifying outbreaks over 90% of the time. Professor Steven Riley, the chief data officer at UKHSA, emphasized the need for innovative disease surveillance methods. He noted that integrating AI with traditional epidemiological practices might help pinpoint sources of foodborne illness outbreaks more accurately.
Though promising, the AI system is still in its early stages and faces challenges in interpreting slang and spelling errors. It struggles with determining the actual source of sickness when reviewers may have misidentified it.
The UK experiences approximately 2.4 million cases of foodborne illness each year, leading to an economic burden exceeding £9 billion. To validate their AI findings, UKHSA researchers manually reviewed over 3,000 cases, demonstrating the system’s potential alongside human analysis.
In conclusion, the study illustrates that integrating publicly available review data with large language models can enhance public health surveillance regarding gastrointestinal illnesses, although results require careful interpretation due to the inherent limitations of relying on such data.