A Bi-lingual chatbot implementation for pandemic response using the transformer-based approach.

Atwine, Mugume Twinamatsiko; Jjingo, Daudi; Nsubuga, Mike; Serunjogi, Richard; Mbabaali, Ibrahim; Lujumba, Ibra; David, Byansi; Kintu, Timothy et al. · PLOS Digit Health · 2026

other · Level V

Where this comes from

Abstract

The COVID-19 pandemic highlighted the critical need for timely and accurate information in effectively managing pandemics. The proliferation of misinformation on public media platforms complicated the management of the pandemic, necessitating a large and constant human resource to meet the demand for trustworthy information. To address this challenge, we developed an intelligent bilingual chatbot that is available 24/7 to provide medically curated up-to-date and approved pandemic management information in English and Luganda. This approach leveraged deep learning to train a chatbot on a growing corpus of curated pandemic-specific information, questions, and answers. Our results demonstrate an implementation of a chatbot that leverages the well-resourced English NLP framework to enable chatting in the Luganda language. They also show that the chatbot is an effective and flexible tool for disseminating accurate information in real-time, while also providing opportunities for continuous improvement through conversation-driven development.