Why mental metaphors do not help us understand chatbot mistakes

Synthese, vol. 207

Abstract

Abstract The function of chatbots like OpenAI’s ChatGPT is based on detecting probabilistic patterns in the training data. This makes them vulnerable to generating factual mistakes in their outputs. Recently, it has become commonplace in philosophical, scientific, and popular discourses to capture such mistakes by metaphors that draw on discourses about the human mind. The two most popular metaphors at present are hallucinating and bullshitting. In this paper, we review, discuss, and criticise these mental metaphors. By applying conceptual metaphor theory, we provide numerous reasons why they do not succeed in providing us with a better understanding of factual chatbot mistakes. We conclude by calling for justifications of the epistemic feasibility and fruitfulness of the metaphors at issue. Furthermore, we raise the question what would be lost if we stopped trying to capture factual chatbot mistakes by mental metaphors.

Authors 2

1. Markus Pantsar Aachen Human Technology Institute Chair for Theory of Science and Technology

   RWTH Aachen University

   Affiliation as printed

   Human Technology Institute, Chair for Theory of Science and Technology, RWTH Aachen University, Theaterplatz 14, 52062, Aachen, Germany

2. Regina E. Fabry corresponding

   Macquarie University

   Affiliation as printed

   Discipline of Philosophy, School of Humanities, Macquarie University, Sydney, NSW, 2109, Australia