This study aims to explore research gaps and suggest future research directions in the field of business administration by comparatively analyzing artificial intelligence (AI) research topics in industry and academia. For this purpose, the analysis subjects include 14,032 AI-related industry articles collected from January 2020 to May 2025, and 1,130 AI-related abstracts selected from 26,086 business administration paper abstracts from the same period. BERTopic-based topic modeling and keyword co-occurrence network analysis were employed as analytical methods. The analysis results revealed, first, that a comparison of topics via BERTopic identified key industry-focused topics that have been relatively overlooked in business administration academia. Second, the keyword co-occurrence network analysis showed that the industry exhibits a hub-and-spoke structure centered on ‘AI’, whereas academia displays a polycentric structure, interpreted as a characteristic of an early research stage. This study proposes specific future research themes for the business administration field by connecting these derived industry-focused topics with the 9 detailed tracks of the Korean Management Review (published by the Korean Academic Society of Business Administration). This research is significant as it empirically analyzes the gap between academia and industry and suggests practical and academic directions for the expansion of AI research in business administration.