AI and Classroom Argumentation: The Chat-With-Cat Project

The Chat-With-Cat project focuses on the development of intelligent tutoring systems to support the learning and teaching of argumentation. The AI tutor provides feedback on argumentative essays written by students in response to controversial questions. The project is grounded in a theoretical and pedagogical framework that integrates the Argumentation Rating Tool (Reznitskaya & Wilkinson, 2017, 2021), Philosophy for Children (Gregory, 2007; Oyler, 2019), the Socratic Challenge (Flammia, 2023), and the Rational Force Model (Næss, 1959), which defines the rational force of a proposition as the product of its acceptability and relevance. The Rational Force Model addresses key limitations of classical structural analysis approaches inspired by Toulmin (1958), which have been criticized for their limited diagnostic power (Chinn et al., 2016; Erduran, 2018). In our project, students’ arguments are analyzed through logical chains of propositions, each classified as present, assumed, or absent. Preliminary results indicate high agreement among human raters and suggest strong potential for automated assessment.