Three wishes, once upon a time, happily ever after: stories have always had a ritualised, repetitive function. AI-generated stories are even more repetitive. We know, by now, the stock style of AI-generated texts: the triads (this, this and this), the not this but that structures, the sycophancy, the overladen metaphors and the lack of conflict. But how does this affect storytelling? After all, humans also use clichés and write bad stories. However, human stories are focalized through a protagonist who experiences something. A story, according to James Phelan’s definition, is more an action than a structure of events: it is “somebody telling somebody else on some occasion and for some purposes something that happened”. With AI there is nosomebody telling. There is no experience to be communicated – there are only words. What does that mean for the reader?

In this talk Jill Walker Rettberg discusses her research on AI-generated stories both from a literary and rhetorical perspective and from the perspective of AI bias. She will analyse examples of AI storytelling in the wild, like the popular AI-generated YouTube videos about billionaires, ghosts and scandal, as well as stories generated specifically to test research hypotheses, and will show how AI-generated stories are influenced both by the technological affordances of machine learning and by the commercial and political content in the models’ training data.  

Jill Walker Rettberg is a Professor of Digital Culture and Co-Director of the Center for Digital Narrative at the University of Bergen, Norway.