ABSTRACT
AutoTutor is an automated computer literacy tutor that participates in a conversation with the student. This paper describes the conversational features that have been incorporated into AutoTutor to improve his performance as a conversational partner. We depict AutoTutor's dialog move options in the Dialog Advancer Network (DAN). The DAN illustrates how AutoTutor adapts his dialog moves to the preceding student turn as well as how AutoTutor gives the floor back to the student. We also discuss how AutoTutor conforms to conversational norms, provides appropriate emotional responses, and appears interested in the student's conversational input.
