The LLM-Ed Framework for Learning Design. Report of the ADMIT project

Within WP4 of the ADMIT project, a structured process has led to the development of the LLM-ed Framework. This task aligns with ADMIT’s objective to promote the inclusive, ethical, and effective use of Large Language Models (LLMs) in teaching and learning. The framework is designed to support teachers in integrating LLMs into the development of learner-centred educational experiences. The initial phase of the framework development involved a comparative analysis of five established Learning Design (LD) models: ABC Learning Design (Laurillard), Carpe Diem (Salmon), Iceberg (OU UK), enABLe (Armellini), and ELDeR (University of Edinburgh). While differing in structure and emphasis, these models share common pedagogical characteristics, including constructivist principles, alignment between learning outcomes, activities, and assessment, a learner-centred orientation, and a collaborative, technology-informed approach to design. From this analysis, several areas were identified where generative AI can contribute, particularly in tasks such as aligning objectives with course content, designing personalised learning pathways, adjusting the tone and complexity of materials, and supporting inclusive design. These areas informed the development of the AI-LD Activity Framework, which presents generative AI as a support tool within the learning design process. The framework was developed through a participatory process that included three dedicated sessions: a joint ADMIT partner meeting (May 15) and two internal meetings hosted by UOC (May 28) and FernUniversität (June 13). These sessions involved academic staff, researchers, administrators, and technical teams in testing preliminary ideas and collaboratively refining the framework.