Marcello Fabio Maida, Associate Professor of Gastroenterology at the Università degli studi di Enna “Kore”, shared a post on LinkedIn about a paper he co-authored with colleagues published in Digestive and Liver Disease:
“Another brick in the wall of AI in Gastroenterology.
Are LLMs equally reliable in every language?
Pleased to share our newly published international study on ChatGPT and colorectal cancer screening across 28 countries and 23 languages.
Our results show overall strong performance, with relevant variability across different languages and regions.
These differences likely reflect training data imbalance, linguistic complexity, and cultural context.
LLMs in medicine must be validated language by language before clinical adoption.
Grateful to the entire WEO World Endoscopy Organization (WEO) Emerging Stars team for this incredible global collaboration: Apostolis Papaefthymiou, Sunil Gupta, Theodor Voiosu, Louis Lau, Stefano Baraldo, Partha Pal, Michael Mwachiro, Zuchelli Tobias, Hugo Uchima, Enrik Aguila, Djilani Bouberra, Helena Degroote, Tolga Düzenli, MD, PhD, Asmaa Gameel, Tsevel Kh, Saraswathi Lakkasani MD, Bayasgalan Luvsandagva, Hasan Maulahela, Nobre Renata, Okubo Yuki, Alessandro Rimondi, Afafe Taiymi, Mostafa Ibrahim.”
Title: Performance of large language models in addressing patient queries on colorectal cancer screening in different languages – An international study across 28 countries
Authors: Maida M, Papaefthymiou A, Gupta S, Voiosu T, Lau LHS, Baraldo S, Pal P, Mwachiro M, Zuchelli T, Uchima H, Aguila EJT, Bouberra D, Degroote H, Düzenli T, Gameel A, Khurelbaatar T, Lakkasani S, Luvsandagva B, Maulahela H, Nobre R, Okubo Y, Rimondi A, Taiymi A, Mostafa I, Conroy G, Dang QDH, Grimaldi J, Hang DV, Heinrich H, Kiryukhin A, Sang Hyub Lee, Legros R, Maas MHJ, Maida CD, Morais R, Pawlak KM, Rath T, Santos-Antunes J, Sudovykh I, Vitello A, Voiosu A
You can read the full article in Digestive and Liver Disease.
