Abstract:

This oral presentation traces the interdisciplinary design and innovation process of Supporting Early chiLdhood education in Blue skills with generative artificial Intelligence (SELBI), a national project funded through the Malta Council for Science and Technology’s Research Excellence Programme. SELBI brings together Early Childhood Education and Care (ECEC), Artificial Intelligence (AI) and marine sustainability to create a retrieval-augmented generation (RAG) platform that supports educators in embedding blue skills and education for sustainability within early learning settings.

Guided by the TPACK framework, the project integrates content knowledge (blue skills), pedagogical knowledge (ECEC practice) and technological knowledge (Generative AI) into a context-sensitive professional learning tool. A five-phase mixed-methods design to Technological Readiness Level 4 includes a nationwide survey with ECEC educators, compilation of a curated knowledge-base repository aligned with local curriculum and policy frameworks, iterative configuration of the RAG model, and a training and validation phase with educators teaching across the ECEC cycle, specifically kindergarten and the first two years of primary education. Prior to conducting trials with the educators, the team conducted internal testing cycles to refine prompts and ensure that system responses were pedagogically robust, goal-oriented, appropriately scaffolded, factually accurate, relevant, clearly expressed, sufficiently complete, and free of hallucinations. These cycles also checked that the platform remained consistent with principles of safety and inclusivity and aligned with relevant educational and institutional policy frameworks.

The presentation foregrounds the collaborative and methodological innovation that shaped SELBI but also the complexities it exposed: reconciling different disciplinary terminologies and priorities; translating topics related to the marine sciences into concepts meaningful for young children; negotiating educators’ mixed levels of digital confidence and ambivalence about AI; and balancing automation with the need to preserve professional judgement and pedagogical agency. It highlights how higher-education cross-sector collaboration between education, marine science, and AI re-framed education for sustainability in ECEC practice and opened new pathways for technology-enhanced professional learning.

The contribution lies in offering a replicable yet adaptable framework for integrating AI-supported sustainability education into ECEC while strengthening teacher education and curriculum development at system level. As the project is in its final stages at the time of the conference, the presentation provides preliminary insights emerging from the design, internal testing, training, and validation phases and reflects critically on lessons learned for future interdisciplinary research and practice.

Keywords:

Early childhood education and care, generative artificial intelligence, blue skills, education for sustainability, professional learning.