Development of Conversational Artificial Intelligence for Pandemic Healthcare Query Support
Wai Lok Woo, Bin Gao, Raid Rafi Omar Al-Nima, Bingo Wing‐Kuen Ling
Abstract
The paper proposes and describes the development of conversational artificial intelligence (AI) agent to support hospital healthcare and COVID-19 queries. The conversational AI agent is called “Akira” and it is developed using deep neural network and natural language processing. It is capable of reading the inputs from the user, understanding the input and identifying the intention, and outputting messages towards the user, and these steps are iterated until the user prompts to exit or the programme is terminated. A deep learning model has been trained, and Akira could converse with the user ranging from the conversation over 7 topics related to COVID-19, common cold and flu, mental health, sexual health, abortions, allergens, drugs and medicine. The paper also describes the importance of designing an interactive human-user interface when dealing with conversational agent. In addition. the context of ethical issues and security concerns when designing the agent has been taken into consideration and discussed. The conversational agent is demonstrated to answer queries from a pool of 57 participants.