Abstract:

Although the“data-driven”paradigm, anchored in big data and machine learning, has significantly expanded the empirical horizons of tourism research, its inherent methodological limitations are becoming increasingly evident. Concurrently, the emergence of Generative AI and Large Language Models (LLMs) marks a transition from“analytical”to“generative”intelligence, catalyzing a methodological revolution in the social sciences. This paper argues that tourism research is undergoing a fundamental paradigm shift: from a“data-driven”approach focused on historical correlations to an “intelligence-generated” paradigm characterized by deep semantic understanding, interactive reasoning, and multi-agent simulation. First, this paper critiques the epistemological bottlenecks of the data-driven paradigm with regard to causal inference and theory building. It then elucidates how the core capabilities of LLMs—contextual awareness, controllable generation, and generative agents—serve as technological cornerstones of the new paradigm. Crucially, this study demonstrates how this shift systematically reconstructs the research lifecycle: inquiry evolves from explaining“facts”to exploring“possibilities”; analysis shifts from mining“correlations”to generating“reasoning”; and theory building transforms from static induction to dynamic“computational experiments.”This transition necessitates a redefinition of the researcher’s role from a data “interpreter” to a “designer”of thought experiments. The study concludes by proposing a forward-looking agenda for navigating the opportunities and ethical challenges of the generative era.

Key words: generative AI, tourism research, research paradigm, data-driven, intelligence-generated

摘要：

以大数据和机器学习为核心的“数据驱动”范式虽极大地提升了旅游研究的经验广度与宏观视野，但其固有的方法论局限日益凸显。与此同时，以大语言模型为代表的生成式人工智能正从“分析智能”向“生成智能”跃迁，为社会科学研究带来根本性变革。本文指出旅游研究正经历一场深刻的范式转型，即从以分析历史数据关联为核心的“数据驱动”范式，转向以深度理解、交互推理和多智能体模拟为特征的“智能生成”新范式。文章首先回溯了“数据驱动”范式在语义理解、因果推断、模拟能力和理论构建上遭遇的“天花板”，随后深入剖析了AI大模型所具备的上下文深度理解、可控生成、交互式推理与生成式智能体四大核心能力，并将其视为新范式的技术基石。研究的核心贡献在于详细阐明了新范式如何对旅游研究的全生命周期进行系统性重塑：研究问题从解释“事实”转向探索“可能性”；数据分析从挖掘“关联”升级为生成“推理”；理论构建从静态归纳演变为动态的“计算实验”；研究成果也从单向的文本输出转变为可交互的“知识系统”。这标志着研究者的角色正从数据的“解释者”转变为思想实验的“设计者”。最后，本文展望了新范式带来的机遇与伦理挑战，并提出一个面向未来的研究议程，旨在推动构建更具前瞻性、批判性和想象力的旅游科学。

关键词: 生成式AI, 旅游研究, 研究范式, 数据驱动, 智能生成