Enhancing Tourist Experience Across Age Groups with Large Language Models: Evaluating ChatGPT as a Virtual Tour Guide

As the smart tourism sector is growing rapidly, large language models (LLMs) have emerged to slowly function as an emerging technology for enhancing the tourist experience. The present study statistically tested the worth of having LLMs as virtual guides in smart tourist scenarios. It comparatively evaluated and analyzed their efficacy and quality in gaining information, user experience, and willingness to use among different age groups. This study recruited 180 tourists in three age categories (Digital Natives, Digital Adopters, and Digital Migrants) via a stratified random sampling method. They were equally assigned to the experimental group and the control group. When compared with conventional tourism digital media, the results showed that virtual tour guides derived from LLMs had tremendous advantages in increasing information acquisition effectiveness, raising interactivity, and optimizing user experience. However, acceptability of LLMs virtual guides was different among people of different ages. Digital Natives and Digital Migrants were more contented and yet willing to accept, while Digital Adopters showed a neutral attitude toward LLMs' profound content. The current study also revealed the key problems of LLMs in intelligent tourism, e.g., information credibility and interactive adaptability. The findings provide theoretical support for digital innovation in smart tourism and practical insights for policymakers, tourism technology companies, and cultural managers.