You can participate via Zoom or join us on site, where the speaker will be present. Welcome!
In this #FrAIday talk, Merel Semeijn engages with fictionalist accounts of verbal human-AI interaction. According to these views, although lay users generally believe that AI systems do not (and cannot) produce meaningful utterances—due to their lack of mental states—they still pretend that AI can, when interacting with it. In doing so, users engage in a kind of “mini-fiction,” treating the AI as if it were an agent with beliefs, desires, and intentions.
Drawing on research in experimental philosophy, Semeijn argues that fictionalism overestimates what laypeople actually believe about AI. Instead, she introduces the notion of "uncaring users"—people who likely don’t know, and more importantly don’t care, whether AI systems produce meaningful utterances. Nevertheless, they behave as if the systems do. This perspective raises new questions about belief formation in human-AI communication.
Merel Semeijn is a postdoctoral researcher in semantics and the philosophy of fiction at the University of Groningen, where she leads the project "Fact, fiction and deception in the digital age." She holds a PhD on fictional discourse, has taught theoretical philosophy, and conducted research in Paris and Leiden. Her work explores how we use language in fiction, lying, and human-AI interaction.