Can't Wait to Learn — Machine-Learning-Diagnosed Adaptive Learning for Refugee and Vulnerable Children in Jordan

Can't Wait to Learn (CWTL) is a self-paced, curriculum-aligned tablet and web game teaching literacy and numeracy, originally built by War Child Holland to reach Syrian refugee children in Jordan's informal tented settlements who had no access to formal schooling. Content, characters and storylines were co-created with Jordanian children to align with the national curriculum.

Extend a proven refugee-education tool from informal settlements into Jordan's mainstream public-school system, reaching both refugee and vulnerable host-community children.

In partnership with Jordan's Ministry of Education (Queen Rania Al Abdullah Centre for Education and Information Technology) and Al-Oun Charitable Society, the programme has expanded into 28 public schools reaching 1,113 children, with a stated ambition to reach up to 3,000 children across northern, central and southern Jordan. The platform's distinguishing feature is a machine-learning diagnostic layer, developed jointly by War Child Holland and Stanford University researchers, that analyses usage-log data to detect 'wheel-spinning' — a child repeatedly failing a skill without making progress — and classify what kind of intervention is needed.

Published research on the ML classifier found it matched expert human judgement on what a stuck child needed, and that roughly half of children experience wheel-spinning at least once during their use of the app; that validation study documents CWTL's Uganda and Lebanon deployments rather than Jordan directly, though the same underlying platform and algorithm run in all country deployments including Jordan's. War Child's own programme reporting for Jordan states that 95% of children who completed both pre- and post-assessments showed significant academic progress, with a 97% improvement in math and 92% in Arabic.

These Jordan figures are self-reported by the implementing organisation rather than drawn from an independent randomised evaluation of the Jordan site specifically — the closest rigorous third-party evaluation, a quasi-experimental study published in the Journal of Development Effectiveness, covers CWTL's Sudan deployment, not Jordan. No gender-disaggregated outcome data for Jordan was found, and CWTL is not a girls-only programme — it serves all vulnerable refugee and host-community children.