Exploring Academics' Orientation Towards Large Language Model Adoption: A Knowledge Management Approach Using <scp>ChatGPT</scp> and Retrieval‐Augmented Generation

Abstract

ABSTRACT Drawing on the Rapid Application Development approach and the Unified Theory of Acceptance and Use of Technology II (UTAUT2), this study develops a generative artificial intelligence (AI) tool for higher education institutions (HEIs), referred to as the Knowledge Management‐Large Language Model (KM‐LLM) application and examines how its actual usage (AU) by academics influences knowledge management (KM) processes. Using Retrieval‐Augmented Generation (RAG) techniques and Large Language Model (LLM) technologies, particularly GPT‐4, the KM‐LLM application was designed to support academic activities within HEIs. Employing a quantitative research approach, survey data were collected from 496 academics working in public and private universities in the Kurdistan Region of Iraq, and the study hypotheses were tested using Partial Least Squares Structural Equation Modelling (PLS‐SEM) with SmartPLS software. The study developed a prototype AI‐based KM application by integrating LLM and RAG technologies. The KM‐LLM application enhances KM processes in HEIs by utilising LLM capabilities, particularly GPT‐4, to create institutional knowledge bases consisting of research outputs and academic studies specific to universities. The findings reveal the level of acceptance of the KM‐LLM application among academics in public and private universities and identify the key UTAUT2 factors influencing behavioural intention (BI) and AU in academic work. This study represents one of the earliest attempts to develop a generative AI application designed specifically for HEIs to enhance KM processes. The significance of this research lies in addressing the limited integration between KM and LLM technologies, as well as the scarcity of studies examining the adoption and usage of generative AI systems among academics.

Ähnliche Arbeiten