Proposal and Development of Tools for Visualizing the Relationship Between Childcare Records and Children’s Qualities and Abilities Using Generative AI
In the National Curriculum Standard for Kindergartens (revised in 2017), it is stated that kindergartens shall endeavor to nurture qualities and abilities. In order to support teachers in understanding children’s development, this paper proposes tools that use generative AI to assess the degree of relevance (relatedness) between observed behaviors of children described in childcare records and qualities and abilities. Multiple methods based on different prompts were designed to calculate the relatedness score for each sentence, and validity was examined by comparing these scores to those assessed by experts. The results for a “foundation of knowledge and skills” and a “basic abilities of thinking, judgement, and expression, etc.” showed high agreement with expert assessments under several methods. Although the overall agreement rate for the “learning dispositions and humanity, etc. to advance towards learning” was low, all methods showed correlation with expert assessments. Furthermore, three visualization methods were proposed, and received a certain level of positive evaluation by teachers and university students, suggesting that they could serve as helpful references for understanding children and grasping the qualities and abilities.
Authors
- Atsuo Hazeyama (ORCID: https://orcid.org/0000-0002-6583-1521)
- Naoki Kato (ORCID: https://orcid.org/0000-0003-4410-4090)
- Shirabe Yuji
Publication Details
- Journal
- Procedia Computer Science
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1016/j.procs.2025.09.293
- Primary Topic
- Diverse Interdisciplinary Research Innovations
- Type
- article
- Field-Weighted Citation Impact
- 0.00