Leveraging chatbots for enhanced decision-making: a comprehensive literature review

Introduction: Chatbots using large language models such as ChatGPT, Google Bard, etc., have become increasingly popular in recent years. The chatbot's ability to simulate conversations with users and process input data and respond based on that information has prompted researchers to investigate the applicability of chatbots in decision-making processes across multiple fields. Current literature has investigated the benefits and challenges of using chatbots in a broad context. This paper presents a systematic literature review of current literature discussing chatbots and decision-making processes, exploring the quality of the decision-making process and user perceptions of using chatbots in the decision-making process. Methods: This SLR aims to provide a comprehensive view of the disciplinary fields in which chatbot decision-making has been evaluated, how chatbots can be used in various fields, and how it compares to human decision-making. Thirty-six articles from seven databases were reviewed in this paper and categorized into six themes: benefits of using chatbot for decision-making, challenges of chatbot-supported decision-making, ethical considerations of using chatbot-supported decision-making, algorithms/tools used in designing chatbots, human vs. AI decision-making, and chatbot decision-making in different fields. Results: An analysis of these themes revealed (i) benefits of personalized recommendations in decision-making, (ii) issues with inconsistency in output, (iii) ethical concerns about chatbots using sensitive information to make decisions, (iv) ChatGPT's decision-making is the most studied, and (v) human vs. AI decision-making. Discussion: The practical and research implications of these findings are further explained in the paper.