Abstract
This chapter urges narratologists and digital humanities scholars to critically (re)examine the narrative affordances of Large Language Models (LLMs) and the parameters underpinning their storytelling dynamics. It takes forward the idea that the latest generation of LLMs such as Gemini and ChatGPT are now able to compose relatively complex narratives, allowing technology companies to promulgate their own narratives about how AI will revolutionise storytelling. It points out that, while AI's literary creativity and aesthetic merits are being debated, its capacity to produce and process formulaic narratives is already being deployed in mundane and bureaucratic scenarios - with far-reaching consequences. Combining insights from narrative theory with natural language processing algorithms, this chapter models a case study to demonstrate the impact of LLMs on the 'banal narratives' found in US immigration processes. It includes as a case study an analytic experiment (involving news data and a corpus of artificial asylum-seeker narratives generated through human-AI interactions) producing new knowledge about the ways LLMs, trained on online discourses, negotiate the affective and emotional collectives forming around critical issues - including the cultural narratives shaping public perceptions of AI.