CUHK · TDLEG 2025-28 | SDGs Integrated

Transformative Experiential Learning:
Generative AI + VR for Disaster Literacy

Adaptive AI avatars · place‑based resilience · formative action for a safer future
GenAI avatars Immersive VR HK hazard regimes Climate action
650+
student beneficiaries
4
disaster VR scenes
A3GIF™
adaptive AI framework
12+
teacher workshops
Principal Supervisor

Dr. Muhammad Sajjad

Research Assistant Professor, Department of Geography and Resource Management, CUHK
muhammadsajjad@cuhk.edu.hk
Expertise: multi‑hazard assessment, extreme weather resilience, experiential learning & AI pedagogy.

Why now? Critical gap in disaster education

Safety + realism gap

Fieldwork limited by extreme weather & logistics. Students miss authentic disaster rehearsal. GenAI+VR fills the void with safe, repeatable plan‑act‑reflect‑retry loops.

False sense of security

HK residents underestimate local landslide/typhoon risk. Place‑based immersion plus adaptive AI avatars builds real decision confidence and risk perception.

Transferable learning

From slope inspection to emergency coordination: students apply skills across hazards, supported by GenAI that calibrates trust & scaffolding.

A3GIF: Adaptive Avatar–Archetype GenAI Immersive Framework
Phase 1 (2026)
Core VR slope scene + learner archetype mapping (PLSPQ). AI prompt refinement & pilot (n≥120). Non‑HMD participation >80%.
Phase 2 (2027)
4 full disaster scenarios: slope, drainage, typhoon shelter, emergency centre. Auto‑archetype personalization, 300+ students.
Phase 3 (2028)
Launch A3GIF, cross‑faculty pilot (English Dept). Open repository, CUHK Teaching Expo, Q1 publications.
Hong Kong–True Immersive Scenes

Slope inspection & Landslide

GEO thresholds, inclinometers, evacuation decisions. Roleplay with Site Supervisor AI avatar.

Urban drainage / flooding

Sensor‑based prediction, real‑time stormwater response (Amber/Red/Black signals).

Aberdeen Typhoon Shelter

Operational management, vessel safety, coordination with emergency crews.

Emergency Ops Centre

Multi‑hazard (typhoon+landslide+flood) decision making, stress & empathy training.

Expected Beneficiaries — Short & Long Term

Short‑term (2026–2028)

  • ≥550 CUHK students: GRMD2221, UGEB2222, UGEB2871, ENGE3950
  • 12–20 faculty members adopting VR/GenAI modules & rubrics
  • 6 immersive HMD workshops + direct disaster simulation training
  • Skill gains: decision‑making, hazard calibration, AI collaboration

Long‑term (2028 onwards)

  • Cross‑disciplinary replication (engineering, urban planning, public health)
  • Open VR scene library + prompt templates → low‑cost adoption across HEIs
  • HK Red Cross Community Resilience Service — scaling citywide disaster literacy
  • Policy influence: evidence‑based disaster preparedness education framework
Aligned with UN SDGs & Global Frameworks
SDG 4 · Quality Education SDG 11 · Sustainable Cities SDG 13 · Climate Action SDG 17 · Partnerships

Global relevance: Supports Sendai Framework for Disaster Risk Reduction (2015-2030) and UNESCO’s ESD for 2030. The project builds climate-resilient communities through AI-enhanced place-based education, directly advancing disaster literacy, risk perception, and inclusive capacity building in the face of increasing extreme weather events.

Key deliverables & projected outcomes
4 VR modules + prompt libraries ≥2 courses adopting GenAI-VR 70% engagement lift | confidence +0.6 cost per student ≤HKD 300 (reuse)
Timeline: Jan 2026 – Apr 2028 Partners: Lingnan U, HKBU, HK Red Cross 600+ total beneficiary count (3 years)

Long-term alignment: CUHK 2026-30 imperatives · Leading AI in T&L · Generative AI ethics · Experiential learning for global challenges