Speaker
Description
In this presentation, we will consider why and how AI-generated feedback can be applied to student submissions in a Moodle database activity.
Research confirms timely feedback is vital for A1–A2 learners (Hattie & Timperley, 2007). Unlike delayed teacher comments, immediate AI feedback prevents cognitive overload and supports "noticing" (Long, 2015). Recent studies (Dai et al., 2023) suggest rapid, scaffolded AI responses can be more pedagogically effective than slow human feedback during early skill development.
Building on this pedagogical need, the presentation introduces an extension of the Moodle Database module using the existing plugin datafield_report, developed by the presenter and sponsored in part by the Moodle Association of Japan (MAJ). The plugin provides a functional expression language—conceptually similar to Excel formulas—that allows teachers to generate secondary data, analytics, and feedback from primary database entries. Central to this system is a generate() function, which accepts an output type (text, image, audio, or video), AI access credentials, a prompt, and an optional file, and then sends this data to an AI provider. The generated output is stored directly back into the Database activity, enabling seamless integration with existing Moodle workflows.
The presentation concludes with a live demonstration of two classroom-ready activities: one that collects student handwriting samples and returns AI-generated feedback with CEFR-aligned ratings, and another that analyzes uploaded video presentations to provide formative feedback. Together, these examples illustrate how the Moodle Database module can be transformed into a flexible platform for scalable, rapid, and pedagogically meaningful AI feedback.
| 発表日の希望 / Preferred Day | いずれの日でも結構です / Either day is fine |
|---|---|
| MAJ R&D Grant | いいえ |
