From Diagnosis to Action: Human-centered AI for Equitable English Support at Scale

This presentation introduces SELA, an AI-assisted, self-diagnostic English assessment embedded in the institutional learning management system. SELA provides a concise snapshot of strengths across reading, listening, speaking, grammar, and vocabulary in a single 60-to-90-minute sitting. Instructors view cohort and class summaries that reveal common needs and inform timely scaffolds such as micro lessons and discipline-specific language support without extra marking. SELA is designed around educational values that elevate learning rather than scoring, strengthening assessment for learning by turning diagnostic evidence into revision tasks, reflection prompts, and goal setting. Equity is supported through accessible delivery, clear criteria, and culturally responsive task contexts. It builds student agency through progress tracking and suggested study pathways and empowers teachers with routines for integrating insights into lesson planning and program review. Participants will leave with adaptable workflows, sample feedback language students can act on, class-integration ideas that connect results to revision cycles and peer review, and communication templates that build trust and transparency about the role of AI. The session aligns with the conference theme by showing how technology, equity, and human-centred innovation transform feedback, reduce friction for educators, and broaden access to effective language support.