AI-Based Simulation of Behavioral Experiments: A Methodological Framework for Hospitality and Tourism Research

This paper proposes and evaluates a methodological framework for using ChatGPT to simulate behavioral experiments in hospitality and tourism research. By generating synthetic traveler samples and prompting them to respond to structured service and choice scenarios, we examine whether AI-generated responses reproduce theoretically meaningful behavioral patterns. Three illustrative simulations are conducted across hospitality decision contexts, including sustainable hotel selection, technology-enabled service delivery, and tiered pricing. Across these applications, ChatGPT-generated data exhibit structured behavioral patterns broadly consistent with prior research, demonstrating the feasibility of AI-based simulation within controlled experimental designs. Rather than positioning generative AI as a substitute for human participants, we propose simulation-based experimentation as a complementary methodological tool for theory development, pre-testing, and exploratory research in hospitality, tourism, and smart tourism. The paper concludes by discussing methodological implications, limitations, and directions for future research on AI-assisted experimental design.