Since 2017, an AI-adaptive tablet numeracy app has taught roughly 250 primary pupils in Dubti, Ethiopia's Afar region, in their own language via a VSO Ethiopia partnership. Unlike onebillion's RCT-evaluated deployments elsewhere, this site publishes no outcome data of its own.
~250 children
Children reached, Dubti/Samara deployment (2017-ongoing)
5.3 months
Additional literacy learning, Malawi RCT (different site, evidence base for the software) (8-month RCT)
3 months
Additional maths progress, UK EEF RCT (different site, evidence base for the software) (EEF-funded RCT)
<10 %
Share of children able to read a word, Tanzania XPRIZE (different site), before (2019 (start))
45 %
Share of children able to read a word, Tanzania XPRIZE (different site), after 15 months (2019, +15 months)
onebillion's onecourse app delivers self-paced, adaptive literacy and numeracy lessons on low-cost tablets, designed to let children progress with minimal teacher input in low-resource classrooms. Since 2017, a deployment in Dubti and Samara — towns in Ethiopia's Afar Region — has taught basic numeracy to primary-age children in Afar, a language rarely served by education technology, through a partnership with VSO Ethiopia.
Activities
Children learn with support from teachers and teaching assistants using the onecourse app on tablets.
Results
onebillion's own project page reports that the Dubti/Samara deployment has reached roughly 250 children, but publishes no pre/post test scores, comparison group or effect size for this specific site. Elsewhere, onebillion's software has been through genuine RCTs: in Malawi, an 8-month RCT with the University of Nottingham found an additional 5.3 months of literacy learning versus a control group, with similar gains for girls and boys; in the UK, an EEF-funded RCT evaluated by the University of Oxford found three months' additional progress in maths; and a 2019 Global Learning XPRIZE field trial in Tanzania raised the share of children able to read a word from under 10% to 45% over 15 months.
Conclusions
None of these rigorous, independently registered studies covers the Ethiopia site, so its own impact remains undemonstrated even though the underlying software has a genuine evidence base elsewhere — real reach in an underserved, minority-language pastoralist region, but without site-specific evidence.
Implementation
Indicative cost
Low (< €50k) — Not disclosed for the Ethiopia site; delivered on low-cost tablets designed for minimal-resource classrooms.
Time to results
Long (> 3 years) — Deployment ongoing since 2017 in Dubti and Samara, Afar Region, Ethiopia.
Staffing & skills
Delivered via partnership with VSO Ethiopia, Children supported by teachers and teaching assistants
Conditions for success
Self-paced tablet software designed for minimal teacher input, suited to low-resource classrooms
Existing proven software base (validated via RCTs in Malawi, UK, Tanzania) reduces technical risk of the Ethiopia deployment
Common failure modes
No pre/post test scores, comparison group or effect size published for the Ethiopia site specifically
Small reach after many years (~250 children since 2017) despite the software's proven effectiveness elsewhere
Where it fits
Governance type
NGO-run deployment via VSO Ethiopia partnership
Scale
small, single-region (~250 children, Dubti/Samara, Afar Region)
Income level
low-income, pastoralist/minority-language region
Commonly funded by
Philanthropic / foundation funding
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Replication kit
Reusable artefacts from this practice — as published by their sources.
STMicroelectronics Foundation, ITU and a blind Milan professor run train-the-trainer digital-literacy classes for blind and visually-impaired adults in Dakar. An …
DepEd's ARAL programme uses Microsoft Reading Progress to automate reading-fluency assessments; a Cabanatuan pilot reduced at-risk readers from 14 to …