Exploring the Drivers of Continued ChatGPT Usage

Generative artificial intelligence (AI) tools such as Chat Generative Pre-trained Transformer (ChatGPT) are rapidly entering university learning, yet hospitality and tourism (H&T) education is distinctive because students must build guest-facing communication and real-time decision-making while using AI responsibly. This study applies an extended Technology Acceptance Model (TAM) to investigate the continued intention of undergraduate students in H&T programme to use ChatGPT. Using structural equation modeling (SEM), it assesses the effects of perceived usefulness (PU), ease of use (PEU), enjoyment (PE), informativeness (PI), prompt engineering skills (PES), and knowledge of ChatGPT (KC) on students’ attitudes (AT) and behavioral intention. Data were collected from 210 undergraduate students enrolled in H&T programs in Hong Kong and Macau. Results show that PU and PEU strongly predict AT and behavioral intention. PE significantly enhances both PU and PEU, while informativeness primarily influences PEU. PES substantially strengthen PU and behavioral intention, whereas KC positively affects all key acceptance constructs. These findings underscore the importance of AI-literacy training and PES development in H&T curricula and suggest institutional strategies to align education with professional expectations in an AI-integrated industry.