生成式AI与人工学术写作的情态对比研究

A Comparative Study of Modality in Generative AI and Human Academic Writing

摘要: 随着生成式人工智能(AI)在学术写作中的应用日益广泛，探讨其与人工写作的语言特征差异具有重要意义。本研究基于系统功能语言学的情态理论，以情态动词为切入点，对比分析AI (DeepSeek)与人工学术写作的情态表达差异。研究发现，人工写作在情态表达上更具灵活性，能够根据语境调整语气，而AI受训练的数据驱动模式，使其语言表达趋于保守，揭示了当前AI写作在学术严谨性与表达多样性上的局限性。

Abstract: With the increasingly widespread application of generative artificial intelligence (AI) in academic writing, exploring the differences in linguistic features between AI and human writing has become of significant importance. Drawing upon the theory of modality in Systemic Functional Linguistics, this study focuses on modal verbs to comparatively analyze the differences in modal expressions between AI (specifically DeepSeek) and human academic writing. The findings indicate that human writing exhibits greater flexibility in modal expression, allowing authors to adjust their tone according to the specific context. In contrast, AI writing, constrained by its data-driven training patterns, tends to be more conservative in its linguistic choices. This contrast reveals the limitations of current AI writing in terms of both academic rigor and expressive diversity.