Treffer: Transforming Reference Services through ChatGPT: Insights from University Libraries in Taiwan.

AbstractThe reference service is a crucial component of university libraries’ services, directly impacting readers. Only high-quality reference services can meet user needs, and their importance is self-evident. This study examines the use of ChatGPT in reference services from the perspective of librarians to understand its potential role. It was found that librarians must improve their competence in emerging technologies, such as AI. Libraries should be cautious in adopting AI, especially given their mission to provide accurate information. While ChatGPT can offer more descriptive responses than traditional chatbots, it still exhibits factual inaccuracies when responding to well-­established queries. It may support basic tasks such as answering simple questions, website navigation, and collection guidance. However, concerns about accuracy and reliability have led some libraries to delay its implementation. The adoption of ChatGPT also involves risks of misuse, ongoing technical support, and funding challenges. In addition to librarian input, it requires collaboration with technical personnel. Rather than a replacement, ChatGPT should be regarded as a supplementary tool, necessitating increased human resources to ensure quality and sustainability. [ABSTRACT FROM AUTHOR]