Toward the Transparent Use of Generative Artificial Intelligence in Academic Articles

With the breakthrough development of generative artificial intelligence (AI), its usage in academic articles is rapidly increasing, and the risk of the lack of research transparency arises with that use. To address this risk, the sources, mechanisms, and quality of AI-generated scholarly content are studied to calibrate our expectations for this technology. The authors find that generative AI has great potential to improve the efficiency of researchers and to enhance research articles but also has significant inherent limitations. Then, they examine the use of generative AI in academic articles from three perspectives: AI-assisted research issue development, AI-assisted addressing of research questions, and AI-assisted research findings communication. On this basis, the authors propose a tiered disclosure strategy based on the generative AI usage context for researchers to transparently use generative AI.