Empowering Doctoral Academic Research: Artificial Intelligence-driven Insights from Large Language Models

Abstract The ever-expanding volume and complexity of academic research pose significant challenges for researchers, particularly doctoral students. In response to these challenges, utilizing Large Language Models (LLMs) has emerged as a promising alternative solution. Such LLMs as ChatGPT, Bing Chat and Google Bard are applied in academic research. This study conducted semi-structured interviews with 50 PhD students and used thematic analysis to explore the application of LLMs in academic research. The results indicate that LLMs assist literature reading by extracting main content, providing research topics, and making reading convenient; assist research design by generating research design ideas; assist academic writing by generating writing ideas, polishing writing, analyzing and visualizing data; assist knowledge construction by offering subject matter expertise and promoting science; assist admin works by writing admin emails. Based on these, a five-dimensional framework of AI-assisted academic research (AIAAR) has been established to explain the assistance of LLMs in academic research. This research not only sheds light on the practical benefits of integrating LLMs in academic research but also provides insights into optimizing their usage for enhanced scholarly productivity and knowledge advancement.