Abstract
This study explores artificial intelligence (AI) applications in menu design optimization within hospitality management. Grounded in choice architecture, arousal and S-O-R theories, it examines how AI-optimized/ menus influence consumer dining behaviors and physiological responses. Based on a baseline menu, ChatGPT enhanced it into three themes: sustainable, healthy, and mixed. A controlled laboratory experiment with 63 participants combined behavioral food choice measures, facial electromyography (EMG), and post-experimental surveys. Random Forest algorithms identified predictive design features. Findings show AI-optimized menus significantly encourage sustainable and healthy choices. EMG-based positive affect mediated the relationship between design and sustainable food selection, but not healthy choices. Machine learning highlighted distinct feature importance: navigation ease and sustainability indicators for sustainability-oriented menus, and health icons for health-focused menus. The study advances hospitality theory by providing empirical evidence of AI’s role in service design and offers practical insights for integrating AI into menu development processes.
