Abstract:

In recent years, MOOCs (Massive Open Online Courses) have been used as a convenient way to learn and acquire new skills, especially in the field of information technologies, which is constantly evolving and expanding while becoming increasingly important in daily life. Therefore, it is important to use automation through the use of modern Information Communication Technologies (ICT) and techniques to improve learner engagement and motivation.

In this study, we opened our popular Experiential Informatics course on the Canvas MOOC platform of Eötvös Loránd University to the general public. It allows practicing teachers to increase motivation with fun ICT tools, students to experience the reach of computer science as a profession, parents to explore apps that facilitate learning, and anyone interested in the wonders of Informatics to learn in a community

To offer a self-paced course, we have integrated chatbot technology that automatically answers basic questions about the course and sets up forums for more complex answers to build a helpful community. For increasing student motivation and engagement, they received feedback through reflections from other users on all major topics, a "like"-based peer review, and a badger-based reward system.

Several publicly available technologies were used to develop the chatbot, including the Facebook Messenger app, wit.ai, and application program interfaces (APIs). Using API technology, the chatbot connected to course content and the Wikipedia Knowledge Center to expand the Conversational Agent's (CA) knowledge base. As for feedback, we tried to involve participants to acknowledge each other's work by asking them to “like” each other’s contributions when appropriate. Five “likes” provided acceptance and completion of a module awarded a badge. Of the 20 modules offered only the first two were compulsory and participants could choose from the others and complete them in any order.

Results:
- Integrating a chatbot with a popular messaging platform like Facebook Messenger improves accessibility and engagement with the chatbot.
- API access to Wikipedia enhances the chatbot's knowledge base and provides a reliable mechanism of communication between the chatbot and the course content, where acquired information depends on design.
- Despite asking questions about the course, participants also made inquiries concerning the personality of the chatbot, implying that the future design of the chatbot should consider the human aspects of these intelligent conversational agents.
- In general, both younger students and adult learners submitted well-thought-out solutions and projects that received sufficient “likes” and thus badges.
- With the use of Badgr, we paid attention to reward only those who properly represented the knowledge they gained.
- Participants highly appreciated the amount and variety of modules (topics) they could choose from and found that the self-paced model met their needs better than a set time frame.

Although we are almost satisfied with the data, we are administering some changes and continuing data collection to compare results with new hypotheses in mind.

Keywords:

MOOCs, Pedagogical Agents, chatbots, CAs, APIs, Facebook Messenger, Badgr, Badge system, Experiential Informatics, Online learning, self-regulated learning.