Page 638 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

2 Use Case Description

2�1 Description

In Gaza, ongoing conflict and infrastructure destruction have led to frequent electricity
blackouts and almost nonexistent internet connectivity. More than 90,000 university students
have lost access to conventional e-learning platforms that require stable connections
and personal computers. The objective of this use case is to deliver an “AI-powered, fully
offline learning environment” that runs on low-power hardware and is accessible from any
smartphone via local Wi-Fi. The innovative approach combines lightweight NLP tutors, AI-driven
content recommendation and automated assessment — all executed on-device — to ensure
uninterrupted, personalized learning during crises.

The solution now uses autonomous micro-agents that communicate over a Wi-Fi mesh. A
knowledge-base agent manages educational content, allowing periodic updates independent
of the AI framework [1]. If any server receives fresh material, it shares it across the mesh so
every hub stays current. English remains the initial language; Phase 2 will integrate Arabic GPT
models for full bilingual support. The design follows the decentralised multi agent coordination
pattern demonstrated in [2], [3], [4].

Expected benefits include 24/7 learning continuity, reduced instructor workload through
auto-grading and scalable deployment to other conflict or disaster zones.

Use Case Status: Seeking partners for scale-up.

Partners

• Al-Aqsa University – Lead institution, deployment & evaluation
• ITU
• Local NGOs & Educational Foundations – Hardware sponsorship, community training
• Open-education content providers – Curated OER materials
• Potential tech partners – Solar and low-power hardware suppliers
• Potential tech partners – Solar and low-power hardware suppliers

2�2 Benefits of use case

• Maintains higher-education access despite infrastructure collapse.
• Demonstrates innovative edge-AI in low-resource settings.
• Level the playing field for students lacking devices or connectivity.

2�3 Future Work

• Expand to lightweight AI apps on Android feature phones.
• Integrate solar-powered community learning hubs across Gaza and neighbouring regions.
• Localize NLP models for Arabic dialects; add speech interface for visually impaired
students.
• Implement federated learning to improve models across multiple campuses securely.
• Seek funding and partnerships to scale pilots from 5,000 to 20,000 students within two
years.
• Integrate Arabic GPT tutor modules after dataset localisation.

602