Purpose

This study, rooted in affordance-actualization theory and communication theory, aims to critically examine how ChatGPT influences users’ transition from new adopters to loyal advocates within the context of travel decision-making. It incorporates constructs including communication quality, personalization, anthropomorphism, cognitive and emotional trust (ET), loyalty and intention to adopt into a comprehensive model.

Design/methodology/approach

This study used quantitative methods to analyze data from 477 respondents, collected online through a self-administered questionnaire by Embrain, a leading market research company in South Korea. Lavaan package within R studio was used for evaluating the measurement model through confirmatory factor analysis and using structural equation modeling to examine the proposed hypotheses.

Findings

The findings reveal a pivotal need for enhancing ChatGPT’s communication quality, particularly in terms of accuracy, currency and understandability. Personalization emerges as a key driver for cognitive trust, while anthropomorphism significantly impacts ET. Interestingly, the study unveils that in the context of travel recommendations, users’ trust in ChatGPT predominantly operates at the cognitive level, significantly impacting loyalty and subsequent adoption intentions.

Practical implications

The findings of this research provide valuable insights for improving Generative AI (GenAI) technology and management practices in travel recommendations.

Originality/value

As one of the few empirical research papers in the burgeoning field of GenAI, this study proposes a highly explanatory model for the process from affordance to actualization in the context of using ChatGPT for travel recommendations.

中文翻译：

---

探索生成式人工智能旅游领域：ChatGPT 对从新用户到忠实采用者演变的影响

目的

这项研究植根于可供性实现理论和传播理论，旨在批判性地研究 ChatGPT 如何影响用户在旅行决策的背景下从新采用者到忠实拥护者的转变。它将沟通质量、个性化、拟人化、认知和情感信任（ET）、忠诚度和采用意图等构建纳入一个综合模型中。

设计/方法论/途径

本研究采用定量方法分析了 477 名受访者的数据，这些数据是通过韩国领先的市场研究公司 Embrain 进行的自填问卷在线收集的。 R studio 中的 Lavaan 包用于通过验证性因子分析来评估测量模型，并使用结构方程模型来检验所提出的假设。

发现

研究结果表明，迫切需要提高 ChatGPT 的通信质量，特别是在准确性、通用性和可理解性方面。个性化成为认知信任的关键驱动因素，而拟人化则对 ET 产生显着影响。有趣的是，该研究表明，在旅行推荐的背景下，用户对 ChatGPT 的信任主要在认知层面上运作，显着影响忠诚度和随后的采用意图。

实际影响

这项研究的结果为改进旅行推荐中的生成人工智能 (GenAI) 技术和管理实践提供了宝贵的见解。

原创性/价值

作为 GenAI 新兴领域中为数不多的实证研究论文之一，这项研究在使用 ChatGPT 进行旅行推荐的背景下，为从可供性到实现的过程提出了一个高度解释性的模型。