Conference Papers

Knowledge Building (KB) is an established learning sciences theory that seeks to promote innovative ideas and idea improvement among students via collaborative engagement in productive discourse. KB discourse supports students to make constructive discourse moves such as questioning, explaining with evidence, adding new information and so on, to advance the collective inquiry. However, current understanding on KB discourse remains limited to students’ online participation. Although small group discussion is a common practice, there is little understanding on the role of verbal discussions to support KB discourse. This paper attempts to address this line of inquiry by assessing student engagement in KB discourse supported by both online and verbal discussions. Data is retrieved from a group of six students in a Grade 6 Social Studies class. The group participated in a 2.5hr lesson designed with opportunities for discussions on the Knowledge Forum (online) and in small groups (verbal). Group talk was transcribed, and Knowledge Forum notes were coded for its semantic level of contribution, with the codes being analysed for weighted connections using Epistemic Network Analysis (ENA). The ENA analysis revealed clear differences in both group and individual engagement between the online and verbal discourse. Notably, students’ contributions on Knowledge Forum showed an apparent pattern of stronger connections among codes of higher semantic levels, suggesting that students were more cognitively engaged in the online discussion than their group verbal talk. Implications for research and practice are discussed.

This study investigated the extent to which levels of certainty impacted the argumentative knowledge construction in individual work and group work. Argumentative knowledge construction has been characterized into simple claims, grounds, qualifiers, counterarguments, and integrated replies to illustrate the components of argumentation and nature of resolving conflicts in argumentation where certainty levels have been divided into uncertain, neutral, and certain. Findings showed that individual and group work differed significantly in terms of levels of certainty for simple arguments and counterarguments. Study implications were discussed

This paper describes the results of a systematic review dealing with the use of personas, avatars, and characters in conjunction with AI-supported tools such as chatbots or generative AI in education. Although the use generative AI in education is gaining traction, this study seeks to systematically review the body of knowledge dealing with personified and conversational approaches to education with both pre-generative and generative AI. The results of the study emphasize the importance of the three key elements of such systems: the use of pedagogical agents, interaction, and personalization. These key elements can be relevant when considering the adoption of the new generation of generative AI in education. Such systems should scaffold learners’ understanding providing guidance and support, promote self-directedness and ensure effectiveness in learning, provide customized learning paths, and promote ethical use.

In recent years, Artificial Intelligence (AI) has significantly increased in digital second language (L2) learning, particularly in supporting vocabulary acquisition. However, research on how AI might facilitate collaborative vocabulary learning is still in its new stage. This study works on investigating the effectiveness of AI recommended contexts in fostering collaborative language learning among young learners. The research employed a self-developed AI-empowered Chinese vocabulary learning system called ARCHe, which was implemented in primary schools in Singapore. A mixed-methods case study approach was conducted with the 2nd-grade students who spoke English as their first language. Preliminary findings indicate that learning Chinese with ARCHe effectively enhances the academic performance of young learners, with the AI-empowered self-generated contexts feature exhibiting a positive impact on collaborative language learning performance. The study offers insights into the integration of AI in digital language learning, with the potential to enhance L2 learning outcomes for young learners.

The logographic nature of Chinese script is a major dissuading factor for learning handwriting. The challenge is the complex psycholinguistic process behind handwriting. Thus, we developed AI-Strokes, a Chinese handwriting learning tool that assists teachers in facilitating students’ handwriting practice in various modalities, and provides personalized feedback for the students. By leveraging a trainable Machine Learning back-end framework, the tool diagnoses and scores students’ handwriting errors. This paper reports a pilot study in a Singapore primary school with an early prototype of AI-Strokes. Two classes of students went through AI-Strokes-based Chinese handwriting lessons (the experimental group) and conventional lessons (the control group) respectively. Pre- and post-tests were administered, and their handwriting processes were analyzed regarding errors in stroke orders, extra/missing strokes, and errors in stroke directions. The results show that the experimental group has yielded significantly better learning gains than the control group. It is posited that the personalized feedback of AI-Strokes has formed a feedback loop to support students’ trial-and-error process in improving their handwriting skills. The multimodal handwriting task design may have also fostered their orthographic awareness through the activation of alternative psycholinguistic pathways during their handwriting lessons.