ABSTRACT
A knowledge gap exists in the application effects of AI-based text mining and ChatGPT as a content analysis tool, which motivates this research. The study utilized four research cases to assess the application effects of AI-powered Text Mining (AITM), compared to the traditional time-consuming content analysis process. The AITM procedure involved four steps: crawling, generating a word cloud, editing codes, and categorizing codes using ChatGPT. AITM in the first case on the daylily tour in Huatan Township found four out of six themes, compared to traditional content analysis. Similarly, the second case on Taiwanese elements in spectator sports yielded two-thirds of the six themes. The third case, which examined the impact of 5G on the travel and tourism industry, demonstrated the effectiveness of automated AI-powered text mining in capturing almost all seven themes. In contrast, the fourth case on travel trends in the metaverse showed that AITM identified seven themes out of nine identified by traditional content analysis. In conclusion, AITM's effects were approximately 78%. This was quite good but still had difficulty identifying themes with subtle and subjective labels, such as authentic experiences, food and agricultural education, nostalgia, art and culture, stakeholders, and triggering factors.
