Designing Custom Chatbots for Classroom Use: Practical Examples, Benefits, and Current Limitations
Abstract
Recent advances in generative AI have made it increasingly possible for individual educators to design and implement their own custom chatbots without advanced programming skills (Abe & Kaneshige, 2025). This workshop introduces practical examples of custom chatbots developed for educational purposes and discusses both their affordances and limitations for classroom use. We would like to share a chatbot creation experience, which we spent one hour building, with a 30-minute presentation. For each chatbot, we will explain its intended purpose, target users, basic specifications, and expected classroom applications. We will also describe the development process, including how the presenters designed the chatbot instructions, refined prompts, and adjusted the chatbot behavior through iterative testing. After introducing these three examples, the workshop will discuss the potential benefits of custom chatbots. In particular, we will show how they can help users access relevant information more efficiently, provide more focused guidance than general-purpose AI tools, and support educators who want to create AI-assisted learning environments tailored to specific instructional goals. The workshop will critically examine current limitations. Even after repeated tuning, custom chatbots may still produce inaccurate, incomplete, or contextually inappropriate responses. Therefore, developers and educators remain responsible for checking the accuracy of the content and deciding whether the chatbot output is appropriate for learners. By combining demonstration, reflection, and critical discussion, this workshop aims to help participants understand how custom chatbots can be used responsibly and realistically in educational settings.
