ChatGPT研究的議題類型與學門類別之關聯分析

An Association Analysis between Types of Research Topics and Academic Disciplines in ChatGPT Thesis

近年來人工智慧中自然語言處理領域快速發展，生成式AI引發了學術界與產業界的廣泛關注。因此本研究以臺灣博碩士論文加值系統為文獻資料取樣來源，分析108篇繁體中文撰寫之ChatGPT相關碩士論文，探討臺灣ChatGPT研究的議題類型與學門類別之關聯。透過專家分群與卡方檢定，研究發現了ChatGPT相關議題可分為：技術開發與實作、學習訓練與其他應用、接受態度與行為偵測。此外，卡方檢定顯示學門與議題間具有顯著的關聯，反映出不同學門對ChatGPT議題的特定偏好，工程學門主要聚焦於技術開發與實作，資訊管理學門偏重接受態度與行為偵測，商業管理及其他學門則集中於學習訓練與應用探索。相關之成果有助於當前ChatGPT議題的系統性梳理，能為後續研究提供探討方向之參考。

In recent years, the rapid development of natural language processing in artificial intelligence, particularly generative AI, has garnered significant attention from academia and industry. This study utilizes the ''National Digital Library of Theses and Dissertations in Taiwan'' as the data source, analyzing 108 ChatGPT-related master’s theses written in Traditional Chinese to explore the correlation between research topics and disciplinary categories in Taiwan. Through expert categorization and chi-square analysis, the study identifies three primary research topics: technology development and implementation, learning and training with other applications, and acceptance and behavior detection. Moreover, the chi-square test reveals a significant association between academic disciplines and research topics, reflecting distinct preferences among disciplines. Engineering disciplines primarily focus on technology development and implementation, information management emphasizes acceptance and behavior detection, and business management along with other disciplines concentrate on learning, training, and application exploration. These findings contribute to a systematic understanding of current ChatGPT-related topics, offering valuable insights for future research directions.