The “Princess Leia” Effect: Maximizing L2 Student Responses to Avatar Chatbots

Abstract

This study reports on the latest stage of a collaborative project at the University of Tennessee (Knoxville, TN, USA). The objective of the project was to design an avatar chatbot (Slabot) capable of eliciting L2 oral responses in Spanish from students and evaluating them according to the criteria established by the American Council on the Teaching of Foreign Languages (ACTFL). After an initial test of Slabot in the spring of 2025 yielded positive results, the present study was conducted in the fall of 2025 in an upper-level undergraduate university course. Of the 19 students who participated, one student’s response reached an aggregate length appropriate for an ACTFL evaluation, and evidence for awarding a particular ACTFL rating is documented. This study concludes that Slabot’s design contributed significantly to the length of this response (24 min 53 s). More broadly, this study suggests that linking an avatar chatbot’s physical features to course content, rather than having students interact with avatar chatbots whose identity is unknown to them, not only facilitates the assignment of an ACTFL ranking but also provides students with an incentive to develop their L2 skills independently.