Application of GAI to Empirical Studies: It’s High Time to Discuss Challenges and Pitfalls

Abstract

ChatGPT became the fastest technological adoption in history within two months of its launch, reaching 100 million users each month. Scholars have been investigating the possible uses of ChatGPT and related tools for peer review and manuscript preparation in academic publishing. The use and effects of GAI tools in academic publishing have been examined in hundreds of thousands of scientific and medical articles. Despite its fastgrowing adaptability, GAI raises multiple concerns. Challenges may include authorship eligibility, hallucination, confidentiality, analysis validation, and precision and recall enhancement. Yet, there are far from sufficient discussions about how information science scholars have been epistemologically, methodologically, technically, and ethically challenged while employing GAI in their research studies. As the obstacles of GAI use in research become clear, there is an urgent need to discuss how to navigate best use practices for Library and Information Science scholars to adopt GAI in research studies. Focusing on potential barriers and pitfalls of GAI use in research studies, this panel will discuss impediments and address solutions in applying GAI to scientific research by sharing specific cases drawing from the panelists’ empirical studies.

Ähnliche Arbeiten