Understanding AI-Generated Experiments in Tourism: Replications Using GPT Simulations
Xiling Xiong, IpKin Anthony Wong, GuoQiong Ivanka Huang, Yixuan Peng
Abstract
The present work explores whether the generative pre-trained transformers (GPT) can complement empirical research in tourism as the GPT extends beyond commercial applications. In particular, we utilized OpenAI’s Python API to interact with the GPT-3.5-turbo. Using GPT as a special subject, we coined AI-generative study (AGS) to validate key findings of 16 scenario-based experiments published in leading journals of tourism and hospitality in two studies. This research contributes to the literature by delineating a new methodology that opens a forum for discussion on alternative means of conducting tourism research. Future studies could also utilize GPT and the ability of generative AI for tourism research in terms of pilot-/pre-testing and cross-validation. In conclusion, we recommend that GPT-generated results should serve primarily as preliminary findings and must be corroborated by data from actual human participants, thus providing converging evidence to support the corresponding research conclusions.