Chatbot FOR TEACHING

Chatbot is an artificial intelligence technology that offers a standardized information management and communication yet providing conversational question-and-answer format and screen tools mimicking human-assistance. Existing chatbot solutions for teaching and learning addresses language learning, economic topics or multiple subject matter, but none is available in Malaysia, what more in Malay language.

CikguAIBot is a chatbot that focuses on assisting the Malay-speaking community in learning the basic concepts and algorithms of Artificial Intelligence (AI). The purpose of the CikguAIBot is to provide an alternative to learning materials and interaction modality with the instructor. Two main topics that are covered in the CikguAIBot which are “Konsep AI” (AI Concepts) and “Algoritma AI” (AI Algorithms). Information given in the chatbot is in multiple formats which are texts, images, videos and links.

The target user of the chatbot ranges from secondary school learners to lifelong learners. CikguAIBot is deployed as a Telegram application and executable through mobile apps and web access. The completion of learning, activities and assessments of the whole content of CikguAIBot takes about one hour.

CikguAIBot is developed using Dialogflow, a development suite on Google Cloud. It has built-in natural language processing features, and integrations with popular communication channels such as Telegram, Whatsapp and Facebook Messenger. The chatbot consists of 65 intents and 7 entities, and is developed using DialogFlow, a Google-based tool. Suggestion chips and cards are used as the interaction means which allow users to navigate from one content to another. Natural language interaction is also allowed so users can chat with the chatbot. Quizzes in the form of true-false and multi-choice questions are created within each topic as a learning reinforcement purpose. Immediate feedback to answers in the quiz is also provided so the students could use the responses as self-learning. The chatbot also offers infographic, links to external resources and videos.

Effectiveness of CikguAIBot is evaluated based on perceived usability, personification, functionality and conversational intelligence; after user’s usage of the chatbot. An online quiz is also prepared to measure the user’s knowledge attainment. Usability testing on CikguAIBot will include usage of content elements such as texts, infographics, videos, web pages and emoticons. Testing on the layout and flow design are also incorporated. Natural language understanding component is also included for both small talk and content enquiries. The chatbot is also capable of understanding the user’s shared emotion and would recommend suitable videos as a response. CikguAIBot preliminary testing has indicated that the chatbot is an interesting learning tool, engaging and offers easy basic knowledge attainment.

Publications:

Nasharuddin, N. A., Sharef, N. M., Mansor, E. I., Samian, N., Murad, M. A. A., Omar, M. K., Arshad, N. I., Shahbodin, F., Marhaban, M. H. (2021), "Designing an Educational Chatbot: A Case Study of CikguAIBot", Fifth International Conference on Information Retrieval and Knowledge Management (CAMP’21).

Sharef, N.M., Murad, M. A. A., Mansor, E. I., Nasharuddin, N. A., Omar, M. K., Samian, N., Arshad, N. I., Ismail, W., Shahbodin, F., (2020), "Learning-Analytics based Intelligent Simulator for Personalised Learning", 2020 International Conference on Advancement in Data Science, E-learning and Information Systems (ICADEIS2020), pp. 1-6, doi: 10.1109/ICADEIS49811.2020.9276858.

Palasundram, K., Sharef, N. M., Nasharuddin, N. A., Kasmiran, K. A., Azman, A., (2019), SEQ2SEQ Models Performance for Education Chatbot, International Journal for Educational Technology, 24(14), pp. 56-68.