Short answer: By pairing AI-augmented, personalized learning with a deliberate shift toward wisdom, creativity, and judgment — not just information delivery — while closing the access gap that determines who benefits from these tools first.

The state of global learning today

Access to school is no longer the world’s binding constraint on education — access to effective learning is. Global youth literacy (ages 15–24) reached 93% in 2024, up from 83% in 1990 [1]. Yet foundational learning has not kept pace with school enrollment. The most recent joint estimate from the World Bank, UNESCO, and UNICEF found that roughly 70% of 10-year-olds in low- and middle-income countries cannot read and understand a simple story — a “learning poverty” rate that jumped sharply after pandemic-era school closures and has not recovered to pre-pandemic levels [2]. Globally, 272 million children and young people remain out of school entirely, a figure that has crept upward rather than down since 2015 [1]. So the challenge is really two challenges stacked on top of each other: get more children into classrooms, and make sure the classrooms they enter actually teach them to read, reason, and think.

AI is now the dominant force reshaping how that gap might close

Artificial intelligence has moved from education pilot programs into national policy. UNESCO is actively working with governments to build AI literacy into school curricula worldwide, and some countries have gone further — the UAE made AI a mandatory subject for all public-school students starting in the 2025–2026 academic year [3]. Personalized tutoring platforms, adaptive assessment, and AI teaching assistants are the fastest-growing category of education technology, though independent analysts disagree sharply on how big that market actually is: 2025–2026 estimates for global AI-in-education spending range from roughly $7 billion to $19 billion depending on what’s counted and whose methodology is used [4]. That spread is itself informative — this is a fast-moving, still-consolidating market, not a settled one, and any single number presented as definitive should be treated with caution.

Beyond tutoring software, longer-horizon efforts to directly augment human cognition are advancing on separate tracks: national brain-mapping projects in the U.S., EU, and China; brain-computer interface research from companies like Neuralink; and early-stage work on smart contact lenses and AR glasses intended to connect learners directly to networked information. None of these are yet mainstream classroom tools, but they represent where augmentation research is heading over the next decade.

The access gap is the real risk, not the technology itself

Every major learning technology in history has reached wealthier, better-connected populations first — print, radio, television, the internet, and now AI are no exception. Left unmanaged, this doesn’t just create an information gap; it risks creating an intelligence-augmentation gap, where some populations gain compounding cognitive advantages years before others get access to the same tools. The pattern has narrowed before — mobile phones and open online courses reached billions of people far faster than earlier technologies did — but it doesn’t narrow automatically. It requires deliberate policy: subsidized access, open-source tools, and curriculum adaptation for local languages and infrastructure.

One concrete proof of concept: the Global Learning XPRIZE, a five-year, $15 million competition that concluded in 2019, challenged teams to build open-source software that could teach children basic literacy and numeracy without a school or a teacher. In a randomized controlled field trial across 141 villages in Tanzania — where most participating children had never attended school — two teams, Kitkit School and onebillion, tied for the win after producing measurable gains in reading, writing, and arithmetic [5]. Their software remains open-source and freely available, but scaling that kind of proven, low-cost tool to the tens of millions of children who could benefit from it remains an open policy problem, not a technology problem.

Intelligence isn’t the same as wisdom

Cognitive science has identified concrete, evidence-based levers for improving learning and brain performance: consistent feedback, sufficient sleep, nutrition, social-emotional support, active reasoning practice, and low-stress learning environments all measurably affect outcomes. But faster information processing is not the same thing as good judgment. This challenge is framed around wisdom deliberately — the ability to weigh competing values, act well under uncertainty, and take responsibility for consequences at scale. Education systems built purely to maximize test scores or content throughput risk producing people who are more capable but no better equipped to navigate the specific global problems — climate change, conflict, inequality, misinformation — that this whole set of Global Challenges exists to address.

Two structural shifts curriculum hasn’t fully absorbed

First, as AI and robotics displace routine cognitive and physical labor, the practical value of memorized content is falling relative to creativity, critical thinking, entrepreneurship, and social-emotional skill — a shift sometimes described as STEM education expanding into STEAM (adding the Arts). Second, learning is decentralizing: free and low-cost online courses, MOOCs, and open educational resources now put a meaningful share of the world’s formal knowledge within reach of anyone with an internet connection. The bottleneck has moved from availability of content to the harder problem of effective use of it — motivation, guidance, and the judgment to know what’s worth learning in the first place.

References

1. UNESCO, International Literacy Day 2025 Factsheet — unesco.org
2. World Bank, UNESCO, UNICEF, FCDO, USAID, Gates Foundation, State of Global Learning Poverty: 2022 Update — worldbank.org
3. Mordor Intelligence, AI in Education Market Size & Trends, citing UAE Ministry of Education policy (May 2025) — mordorintelligence.com
4. Market range compiled from Mordor Intelligence and Grand View Research industry reports, 2025–2026 — mordorintelligence.com / grandviewresearch.com
5. XPRIZE Foundation, $15M Global Learning XPRIZE Culminates With Two Grand Prize Winners (2019) — xprize.org

• Make increasing individual and collective intelligence an explicit national objective of education policy, not just a byproduct of raising test scores or graduation rates.
• Expand online, self-paced, lifelong learning infrastructure to serve aging populations and workers facing repeated career transitions as automation reshapes labor markets.
• Increase public R&D funding for human-AI collaborative learning tools, with independent evaluation of learning outcomes rather than reliance on vendor claims.
• Deploy AI-driven personalized learning tools around evidence-based factors — a learner’s pace, prior knowledge, retrieval practice, and spaced repetition — rather than around “learning styles,” a framework cognitive science has not been able to validate despite its popularity in commercial edtech marketing.
• Broaden STEM education into STEAM, integrating the arts and humanities to build creativity, ethical reasoning, and social-emotional skills alongside technical literacy — capacities that remain hard to automate.
• Shift curriculum design from single-profession mastery toward transferable skill combinations, reflecting how frequently workers now change roles and industries over a career.
• Ground teacher training in current cognitive- and educational-neuroscience findings — for example, evidence on the role of sleep, spaced retrieval, and stress in memory consolidation — rather than in outdated or unvalidated pedagogical theories.
• Study contrasting national models directly: Finland’s interdisciplinary, phenomenon-based teaching and South Korea’s intensive, technology-integrated system both post strong outcomes through very different means, and comparing them can inform policy elsewhere.
• Build on the open-source tools produced by the Global Learning XPRIZE (concluded 2019) — freely available literacy and numeracy software proven in a randomized field trial — instead of starting from scratch on early-literacy technology.
• Prioritize closing the intelligence-augmentation gap directly, through subsidized access, open licensing, and infrastructure investment, so that AI-era learning tools don’t repeat the decades-long lag in adoption that earlier technologies showed between wealthy and low-income regions.

Sub-Saharan Africa: The region has posted real gains — adult literacy rose from 65% to 69% between 2015 and 2024, among the fastest improvement rates of any region — but population growth means the absolute number of illiterate adults still grew, from 196 million to 225 million over the same period [1]. More than half of the world’s 272 million out-of-school children, adolescents, and youth live in Sub-Saharan Africa, and low-income countries region-wide have an out-of-school rate of 36%, compared with 3% in high-income countries [1]. Closing this gap is arithmetically central to the global education challenge: progress here moves the global average more than progress anywhere else.

Middle East and North Africa: According to UNICEF’s 2025 regional analysis, at least 30 million children are out of school across 12 MENA countries — roughly 1 in 3 — concentrated in countries affected by conflict and fragility [2]. The deeper problem is foundational skills even among children who are enrolled: 6 in 10 children in the region cannot read and understand a simple text by age 10, despite relatively high primary completion rates [2]. That combination — decent enrollment, weak foundational learning — is the region’s central education challenge.

Asia and Oceania: The region shows the widest internal variation of any Global Challenge 9 region. Central and Southern Asian youth literacy rose sharply, from 87% to 94% between 2015 and 2024, while adult literacy in the same subregion climbed from 72% to 77% [3]. Sub-Saharan Africa aside, Central and Southern Asia still holds one of the largest concentrations of adults without basic literacy in absolute terms, given population size. Oceania (excluding Australia and New Zealand) has seen comparatively little movement, with youth literacy holding around 73% [3]. South Korea remains a global outlier for household investment in supplementary private education, a long-documented feature of its system reported annually by Statistics Korea (KOSTAT) [4].

Europe: The EU reached 45% tertiary education attainment among 25–34-year-olds in 2025, matching the bloc’s own 2030 target five years early [5] — though Eurostat’s regional breakdown shows this masks sharp internal disparity: only about 35% of EU regions have individually hit that benchmark, with rural and peripheral regions lagging well behind metropolitan centers [6]. The EU had 19.1 million tertiary students in 2024 [7]. The bloc’s Digital Education Action Plan (2021–2027) remains the active policy framework guiding digital and AI integration in schools across member states [8].

Latin America and the Caribbean: The region entered the pandemic already in a learning crisis — UNICEF’s pre-pandemic baseline found more than a third of students below minimum reading proficiency and just over half below minimum math proficiency [9] — and COVID-era school closures made it dramatically worse. A 2022 joint World Bank/UNICEF/UNESCO report found four in five sixth-graders in the region were expected to lack basic reading comprehension, with the pandemic setting learning outcomes back by more than a decade in some estimates [10]. That report has not been superseded by a more recent joint regional assessment as of this writing, so it remains the primary-source benchmark for the region’s current learning crisis, even though the underlying data reflects acute pandemic-era disruption rather than a stabilized post-recovery picture.

North America: In the U.S., total student loan debt — federal and private combined — has grown past $1.86 trillion as of early 2026, per Federal Reserve data, now the second-largest category of household debt after mortgages [11]. On the international side, the U.S. hosted a record 1,177,766 international students in the 2024/25 academic year, a 5% year-over-year increase, according to the Institute of International Education’s Open Doors 2025 report [12] — though IIE’s more recent Fall 2025 snapshot of participating institutions shows a 1% enrollment decline for fall 2025, driven by a sharp drop in new graduate enrollments, a trend worth watching rather than treating as settled [12]. In Canada, the absence of a federal ministry of education means curriculum and credentialing decisions remain provincial, producing persistent variation in how university output aligns with regional labor-market needs.

References (regional section)

1. UNESCO, International Literacy Day 2025 Factsheet — unesco.org
2. UNICEF, At least 30 million children out of school in the Middle East and North Africa (2025) — unicef.org
3. United Nations Statistics Division, SDG Indicators Report 2025, Goal 4 — unstats.un.org
4. Statistics Korea (KOSTAT), private education expenditure survey — kostat.go.kr (figure needs a current-year pull before publishing; flagged previously as stale)
5. Eurostat, Educational attainment statistics — ec.europa.eu/eurostat
6. Eurostat, 35% of EU regions met the EU’s higher education target (2025) — ec.europa.eu/eurostat
7. Eurostat, Tertiary education statistics — ec.europa.eu/eurostat
8. European Commission, Digital Education Action Plan (2021–2027) — education.ec.europa.eu
9. UNICEF, Reimagining Education in Latin America and the Caribbean — unicef.org
10. World Bank, UNICEF, UNESCO, Four in Five Children in Latin America and the Caribbean Will Not Be Able to Understand a Simple Text (2022) — worldbank.org
11. Federal Reserve, G.19 Consumer Credit release — federalreserve.gov
12. Institute of International Education, Open Doors 2025 Report — iie.org

Global Challenge 9 Video

Editorial note: This video reflects an earlier version of Global Challenge 9 and references two figures that have since been updated in the sections above. First, the AI-in-education and e-learning market projections cited in the video ($1 billion in 2018 growing to $6 billion by 2025, $30 billion by 2032, and e-learning surpassing $370 billion by 2026) come from a single commercial forecast; current independent market-research estimates vary widely, so the Overview above reports policy and investment trends from primary sources rather than a specific dollar figure. Second, the video’s reference to AI assistants matching instruction to “unique learning styles” reflects a framework that cognitive science has not been able to validate; the Actions section above reframes this around evidence-based personalization instead. A revised video is planned; until then, readers should treat the Overview and Actions sections as the current, authoritative version of this conten

Video Transcript: Global Challenge 9 — Education and Learning

How can education and learning make humanity more intelligent, knowledgeable, and wise enough to address its global challenges?

Artificial intelligence is poised to significantly enhance human intelligence by diagnosing optimal learning methods and content tailored to individual needs — like how glasses improve vision. AI and related technologies will augment human cognitive capabilities, potentially creating “augmented geniuses.” Companies are exploring smart contact lenses and augmented-reality glasses connected to the Internet of Things (IoT), which could accelerate learning, minimize miscommunications, and make education more engaging.

AI in education is expected to grow from $1 billion in 2018 to over $6 billion by 2025, and $30 billion by 2032, while e-learning is projected to surpass $370 billion by 2026. The integration of AI and massive online educational resources is revolutionizing learning, supported by initiatives like UNESCO’s collaboration with curriculum developers to incorporate AI into all schools. As the education market expands, the cost of technologies and learning designs is expected to lower, making them more universally accessible.

Innovative educational practices are emerging globally. Finland’s interdisciplinary approach teaches events and phenomena rather than traditional subjects. China plans to equip its 400,000 elementary schools with 3D printers within two years. South Korea uses telepresence robots with remote native speakers to teach English, while Dubai employs 3D glasses, holograms, and virtual reality (VR) for immersive learning experiences.

Given the increasing power of technological capacities available to individuals, greater emphasis must be placed on ethics, values, and civic responsibilities. Enhancing education tends to strengthen democratic processes.

What actions will help address this challenge?

• Make increasing individual and collective intelligence a national goal, integrating advances in cognitive science and brain research into education systems.
• Encourage online lifelong learning in anticipation of aging societies and technological change.
• Invest in research and development for AI-human symbiotic evolution to enhance learning and cognitive abilities.
• Develop personal AI assistants that can identify and address individual learning needs in harmony with unique learning styles.

To learn more about education and learning, visit The Millennium Project website at millennium-project.org to see how this fits into the global situation and future possibilities. See State of the Future 20.0, available on The Millennium Project website.