自2023年初生成式人工智能流行以来，人们对生成式人工智能的评价褒贬不一。在 TED 这个具有广泛影响力的平台上，生成式人工智能成为被演讲者广泛讨论的话题，为研究人们对 AI 的态度提供了资源。本文以马丁的评价理论为基本框架,运用定量研究和定性分析的方法, 对 Ted Talk 上关于生成式人工智能的演讲进行话语分析,研究演讲者是如何运用评价理论中的态度、介入和级差资源，实现演讲语篇的人际功能，研究人们对生成式人工智能的态度。研究发现语篇中级差资源占比最多, 态度和介入资源占比较小。同时发现，大多数演讲者对生成式人工智能持有积极态度，呼吁大家接受、拥抱人工智能，在人工智能的协助下提升自己的效率。通过分析演讲者的态度和评价，揭示了围绕人工智能发展的话语，同时提供了对语言特征和社会视角的实用见解。

The surge in popularity of generative AI, especially since the beginning of 2023 with the influential platform of TED Talks, provides a unique opportunity to examine how advanced AI concepts are discussed and embraced globally. Using Appraisal Theory (AT) as the basic framework, employing both quantitative research and qualitative analysis methods, this study conducts discourse analysis on TED Talks regarding generative artificial intelligence. It investigates how speakers utilize attitude, engagement, and graduation resources from the AT to achieve interpersonal functions in their speeches. The study finds that graduation resources dominate the discourse, while attitude and engagement resources are less prominent. Furthermore, it observes that the majority of speakers hold a positive attitude towards generative artificial intelligence, advocating for acceptance and embracement of AI to enhance efficiency with its assistance. By analyzing the speakers’ attitudes and evaluations, this thesis reveals the discourse surrounding the development of artificial intelligence, offering practical insights into linguistic characteristics and social perspectives.