The article argues that literary disciplines provide indispensable tools for building culturally literate AI, addressing the limitations of current LLMs which are described as massive, automated, and monolingual.
It develops a layered framework to create more nuanced textual models and pluralistic interpretations of AI by connecting critical theory debates with structural monolingualism. The essay suggests applying world literature approaches to global AI textuality through concepts of macrostructure, circulation, and untranslatability.
This approach emphasizes the natural intersections of literature and AI development to negotiate cultural struggles inherent in automated textual analysis.