随着人工智能和大数据技术的飞速发展，大语言模型，如大语言模型等已成为全球范 围内的热点技术。这些技术在教育、商业、娱乐等多个领域展现出巨大的应用潜力。

  本文通过深入访谈和问卷调查的方式，探究了北京师范大学韩国留学生对大语言模型 的认知与应用情况，以 ChatGPT 为代表的大语言模型展现了高度的用户认可度与使用频 率，尤其是在学术领域。研究发现，留学生群体普遍认识到这类技术带来的正面影响，但 也表达了对其局限性及社会效应的关注，特别是涉及教育、个人成长、隐私保护和伦理道 德方面。

  本研究重申了技术接受模型（TAM）在大语言模型场景中的有效性，强调了技术的实 用性和易用性对用户接受意愿的正向驱动作用。进一步分析揭示，个体的创新性、自我效 能感、以及对技术的信任感与其使用大语言模型的意愿密切相关。同时，文化价值观如集 体主义、不确定性规避和长期导向，被发现能调节留学生对技术的接受度，影响他们对技 术的评价和采用决策。

  研究结果对教育实践、技术设计及跨文化技术传播具有重要启示，强调在推进人工智 能技术应用的同时，需关注用户多样性，提升技术的人文关怀，加强数据安全与隐私保护， 并在设计中融入文化敏感性。然而，研究受限于特定样本群体，且主要依赖相关性分析， 未来研究可进一步拓展样本范围，采用更复杂的设计来探究因果关系，深化对大语言模型 社会影响的理解。

With the rapid advancement of artificial intelligence and big data technologies, large language models have become a globally recognized technological frontier. These technologies demonstrate vast potential applications across multiple domains, including education, business, and entertainment.

This study, employing in-depth interviews and surveys, examines the perceptions and applications of large language models among South Korean students at Beijing Normal University, focusing on ChatGPT as a representative example. It reveals high levels of user acceptance and frequency of use, particularly in academic settings. The research finds that while students recognize the positive impacts of such technologies, they also express concerns about their limitations and social effects, especially regarding education, personal development, privacy protection, and ethical issues.

The study reaffirms the effectiveness of the Technology Acceptance Model (TAM) in the context of large language models, highlighting the positive influence of perceived usefulness and ease of use on user acceptance. Further analysis shows that individual traits such as innovativeness, self-efficacy, and trust in technology are closely linked to the willingness to use large language models. Additionally, cultural values, including collectivism, uncertainty avoidance, and long term orientation, are found to moderate the students' acceptance of technology, influencing their assessments and adoption decisions.

The findings carry significant implications for educational practices, technology design, and cross-cultural technology dissemination. They underscore the need to consider user diversity in advancing artificial intelligence applications, to enhance humanitarian concern in technology, to bolster data security and privacy, and to incorporate cultural sensitivity into design. However, the study is limited to a specific demographic group and primarily relies on correlational analyses. Future research could expand the sample size and employ more complex designs to explore causal relationships, thereby deepening our understanding of the societal impacts of large language models.