This paper introduces the concept of AI-textuality that extends Bakhtin’s notion of intertextuality to encompass interactions involving texts produced by generative artificial intelligence (GenAI). Intertextuality provides a valuable lens for understanding how GenAI outputs are created through the assemblage of digital and multimodal texts from vast datasets. Building on this perspective, this paper presents a model to theorize the intertextual nature of AI-generated texts in educational settings. The model frames human-AI interaction as an interplay between humans’ experiences in a material space and GenAI’s computational processing in a digital space. Drawing on data from a high school classroom, the model illustrates how students engaged with GenAI to discuss key ideas in the lesson. It highlights how students’ engagement with AI-generated texts was shaped by their textualization of experiences, the intertextual probability of GenAI’s responses, and the joint construction of knowledge between students and GenAI. Through the illustration, the paper argues that AI-textuality has the potential to revise and expand traditional understandings of intertextuality to incorporate our interactions with GenAI in the digital age.

中文翻译：

---

AI 文本性：扩展互文性，将人与 AI 的交互与生成式人工智能的交互理论化


本文介绍了 AI 文本性的概念，该概念扩展了 Bakhtin 的互文性概念，以涵盖涉及生成式人工智能 （GenAI） 生成的文本的交互。互文性为理解 GenAI 如何通过从大量数据集中组合数字和多模态文本来创建 GenAI 输出提供了一个有价值的视角。基于这一观点，本文提出了一个模型，将教育环境中 AI 生成文本的互文性质理论化。该模型将人类与 AI 的交互定义为人类在物质空间中的体验与 GenAI 在数字空间中的计算处理之间的相互作用。该模型利用高中课堂的数据，说明了学生如何与 GenAI 互动以讨论课程中的关键思想。它强调了学生与 AI 生成文本的互动如何受到他们对体验的文本化、GenAI 回答的互文概率以及学生和 GenAI 之间的知识联合构建的影响。通过这个插图，本文认为 AI 文本性有可能修改和扩展对互文性的传统理解，以将我们与 GenAI 的互动纳入数字时代。