Quality Education for All
School access has grown, yet only 58% of primary pupils reach minimum reading proficiency: high-level education for everyone is unsolved.
Open in the interactive tree →Compulsory schooling and literacy campaigns raised enrollment worldwide, but learning lags: many children sit in school without learning to read. Poor, rural and female learners are hit hardest. Higher education (secondary, tertiary, lifelong) is far more unequal than basic literacy.
As of October 2026
The 2026 UNESCO Global Education Monitoring Report counts 273 million children and youth out of school (17%, unchanged since 2015; 3% in rich vs 36% in low-income countries). Only 58% reach minimum reading proficiency and 44% in math at the end of primary school; in low-income countries only 13 youths from the poorest quintile complete upper secondary per 100 from the richest. Aid to education fell 8% between 2023 and 2024, and the share of academically qualified primary teachers in low-income countries dropped from 89% (2013) to 78% (2024).
What is missing
- Enough qualified teachers
- Reliable funding (aid to education fell 8% in 2023-24)
- Instruction in learners' mother tongues with matching materials
- Better data on actual learning (large gaps in SDG 4 monitoring)
- Cheap digital learning aids that work without stable internet
Becomes possible once solved
- A broad skilled workforce in every region
- Informed participation in democracy and economy
- More researchers and founders worldwide
Open steps
- Tutoring that makes learning last Medium AI leverageDesign AI tutors that raise lasting learning at scale and avoid the crutch effect seen without guardrails.
- Materials in mother tongues Medium AI leverageLessons, exercises and tests in learners' mother tongues, including languages with little digital text.
- Tutoring without stable internet Medium AI leverageTutoring that works on cheap phones and shared devices with little or no connection.
- Measuring actual learning Medium AI leverageMeasure what pupils actually learn where assessment data are missing, without costly national tests.
- Tools that extend scarce teachers Medium AI leverageTools that cut teacher workload and extend scarce qualified teachers, tested in real classrooms.
Where AI could help
Medium AI leverage. Tutoring trials show real learning gains, but teachers, funding, devices and local-language content, not software, are the main gaps.
- Personal tutoring that adapts to each pupil, with a teacher supervising
- Generating lessons, exercises and tests in learners' mother tongues
- Low-bandwidth or offline tutoring on phones and shared devices
- Assessing actual learning from written and spoken answers where test data is missing
Shown so far
- In June 2025 Scientific Reports published a randomized crossover trial with 194 Harvard physics students in which a custom AI tutor produced significantly more learning in less time than an in-class active-learning lesson. source
- In June and July 2024 a World Bank randomized trial in Nigerian secondary schools found a six-week, teacher-supervised GPT-4 after-school English program raised test scores by about 0.3 standard deviations. source