Learning through teaching robots has been shown in some studies to produce learning effects similar to those of teaching humans. However, it cannot be said that these learning tools, known as teachable agents, are effective for all users. This is because most of the studies on teachable agents have been conducted on younger children who tend to prefer to take care of others. We hypothesized that if adult subjects interacted with these agents, then their learning experiences would differ based on their individual personalities. Therefore, in this study, we developed a novel-style teachable chatbot and investigated the relationship between the evaluation of learning using this system and personality.