2027 GEM Report consultation

Share your inputs with us!

Research has consistently demonstrated that school meal programmes can have a positive impact on learning outcomes through improved attendance, concentration, retention, and overall student well-being. Given the focus of the upcoming GEM Report on quality and learning, I wonder whether this relationship could be explored or reflected within the report's analytical framework.
In addition, WFP is currently collaborating with UNESCO on the development of a new SDG 4 indicator on school meal coverage. Do you see potential for this indicator to strengthen monitoring of factors that contribute to learning outcomes? Could it also be considered as part of the evidence base or monitoring framework discussed in the next GEM Report?

Response to the consultation on the concept note for the 2027 GEM Report
Melanie Ehren and Sheena Bell

Melanie Ehren; Vrije Universiteit Amsterdam, the Netherlands; m.c.m.ehren@vu.nl
Sheena Bell; University of Toronto, Canada; sheena.bell@utoronto.ca

The concept note makes a timely and valuable contribution by shifting attention from universal global targets towards ambitious but realistic national benchmarks. We particularly welcome its intention to incorporate national definitions of education quality and to support policy learning about why education systems progress at different rates. Our comments are intended to strengthen these ambitions. We offer two related suggestions: broaden the conceptual framework of education quality, and adopt a more dynamic account of system change and learning improvement.

1. Broaden the conceptualisation of education quality
   The proposed framework includes important concepts such as learner agency, professional learning communities and well-being. Nevertheless, its organisation around inputs, processes and outcomes, its emphasis on individual learning outcomes in reading, mathematics, environmental science and IT skills, and the aim to isolate factors and mechanisms at the learner-, teacher-, school- and system-level that contribute to these primarily reflect a human-capital perspective grounded in school-effectiveness research. This is a legitimate and useful perspective; it is however just one perspective of education quality and, on its own cannot fully realise the concept note’s broader ambition to reflect national definitions of quality and education’s contribution to collective well-being and sustainable relationships with the planet.

In our forthcoming chapter, ‘Conceptualizations of Education Quality in Relation to System Transformation’ , we synthesize philosophical perspectives on what education is for, whom it should benefit and who should define quality. Drawing on Labaree (1997, 2011), Tikly and Barrett (2011), Baxen et al. (2014), Brighouse and colleagues (2016, 2018), Biesta (2020), and Akyeampong and Higgins (2025), we distinguish three broad paradigms that, together, offer a broader conceptualization of education quality. These three conceptualisatons: human capital, human rights and human capabilities together allow for diverse national definitions of quality to be taken into account into the GEM report. Table 1 summarizes these.

Table 1. Conceptualisations of education quality (attached image)

These different conceptualizations of education quality also imply different indicators to measure progress. When UNESCO has the aim of incorporating national definitions of quality into its monitoring and GEM report, we argue that the selection of country case studies and measures of progress need to be selected by the various conceptualizations of quality; rather than the single set of learning outcomes and input/process indicators that are now put forward.

The concept note could make the underlying conceptions of quality more explicit and visible and examine whether countries define and pursue quality primarily in human-capital, human-rights or human-capabilities terms, or through combinations of these. It could also point to tensions between these different approaches to education quality and the values and purposes they ascribe to, and implications for system leaders, policy designers, or implementors at the middle tier or school-level actors. This would align the analytical framework more closely with SDG 4’s expansive view of education and prevent national variation from being interpreted solely through a common set of learning-outcome and input/process indicators.

1. Treat system change as dynamic, uneven and context-dependent
   The concept note’s interest in countries that have improved fastest offers a clear organising question. At the same time, system change and learning improvement are rarely linear and may even be at odds. International assessments often show that countries fluctuate in reported student outcomes, and aggregate progress can coexist with declining outcomes or widening inequalities for particular regions, schools or learner groups. Improving instructional quality across an education system requires changing the underlying practices, norms and beliefs of teachers and subnational support actors, which is inherently a slow process. Selecting countries mainly because they show rapid national improvement may therefore obscure the interactions and distributional patterns through which change occurs.

Authors in the forthcoming Oxford Handbook argue that there is little research to support an image of “transformation as reinvention”; reviews and analyses of research on education policy, innovation, and change in the US and globally rather suggest a different image; one that is described by the authors as “transformation as evolution” – a gradual, uneven, yet sustaining metamorphosis within educational ecologies (Datnow et al., 2022a; Park et al., 2025; Peurach et al., 2019; Peurach et al., 2022).

From this perspective, the concept note’s question about trends in learning achievement (p. 6) provides a particularly productive entry point. We suggest extending the methodology to specify how trajectories, discontinuities and interactions among system levels (classroom, school, district, national) will be examined over time. This would enable the report to identify not only whether outcomes changed, but also for whom, where, through which mechanisms and under what contextual conditions. In line with our previous point, the selection of cases should focus on selecting and identifying countries and policies that represent different conceptions of quality and understanding how they manage and improve learning and teaching accordingly. For example, some countries are orienting their curricula and resource allocation around improving skills for employability (e.g., Rwanda’s aim to achieve middle income status through major expansion of secondary skills and TVET) (MINEDUC, 2022). Other systems are shifting their systems with reforms aiming at greater local, cultural relevance. These include Zimbabwe’s inclusion of heritage based curricula and ubuntu in its curricula, or Senegal’s MOHEBS reform to develop curricula and language of instruction in 6 national languages (Bell, 2025; MOPSE, 2026). Case selection could consequently combine cross-national benchmarking with within-country analysis. National benchmarks can provide a first phase for identifying contrasting regional or local trajectories, including ‘positive outliers’ that achieve progress under conditions comparable to those faced elsewhere in the same system. Such cases may offer more actionable lessons than comparisons with countries whose institutional and social contexts differ substantially. They can also reveal how different conceptions of quality shape the management and improvement of teaching and learning in practice.

Taken together, these two suggestions would reinforce the report’s distinctive contribution. A plural conception of quality would clarify what counts as progress, while a dynamic systems perspective would strengthen explanations of how progress occurs. This combination would support more context-sensitive national target-setting and richer policy learning for the post-2030 education agenda.

References
Akyeampong, K., and Higgins, S. (2025). Chapter 9. Engaging Parents, Extended Families and Communities; Second Chance Programmes in Conflict-affected Liberia (p.242-260). In: K. Akyeampong, and S. Higgins (Eds.). Reconceptualising the Learning Crisis in Africa; Multi-dimensional Pedagogies of Accelerated Learning Programmes. London: Routledge
Barrett, A., Chawla-Duggan, R., Lowe, J., Nikel, J. & Ukpo, E. (2006). Review of the ‘international’ literature on the concept of quality in education, EdQual. A Research Consortium on Implementing Education Quality in Low Income Countries. https://www.edqual.org/publications/workingpaper/edqualwp3.pdf/at_download/file.pdf
Baxen, J., Nsubuga, Y., & Johanson Botha, L. (2014). A capabilities perspective on education quality: Implications for foundation phase teacher education programme design. Perspectives in Education, 32(4), 93-105.
Biesta, G. (2020). What constitutes the good of education? Reflections on the possibility of educational critique. Educational Philosophy and Theory, 52:10, 1023-1027, DOI: 10.1080/00131857.2020.1723468
Bell, S. (2025). Strengthening the Middle Tier for Foundational Learning: School Support, Data Use, and Policy Implementation in Senegal (Science of Teaching, p. 48). RTI-International. https://scienceofteaching.site/wp-content/uploads/2025/08/Senegal-Middle-Tier-Study-Report.pdf
Brighouse, H., Ladd, H. F., Loeb, S., & Swift, A. (2016). Educational goods and values: A framework for decision makers. Theory and Research in Education, 14(1), 3-25.
Brighouse, H., Ladd, H. F., Loeb, S., & Swift, A. (2018). Educational goods: Values, evidence, and decision-making. University of Chicago Press.
Brock-Utne, B. (2016). The ubuntu paradigm in curriculum work, language of instruction and assessment. International Review of Education, 62(1), 29-44.
Labaree, D. F. (1997). Public goods, private goods: The American struggle over educational goals. American educational research journal, 34(1), 39-81.
Labaree, D. F. (2011). Consuming the public school. Educational theory, 61(4), 381-394.
MINEDUC. (2022, November 16). Education stakeholders pledged to intensify learning outcomes. Rwanda Ministry of Education. https://www.mineduc.gov.rw/news-detail/education-stakeholders-pledged-to-continue-improving-learning-outcomes
MOPSE. (2026). Heritage Based Curriculum Syllabi. Ministry of Primary and Secondary Education. https://www.mopse.gov.zw/heritage-based-curriculum-syllabi/
Peurach, D. J., Datnow, A., Park, V., & Spillane, J. S. (in prep). The Oxford handbook on systems transformation in education. Oxford University Press.
Tikly, L., & Barrett, A. (2011). Social justice, capabilities and the quality of education in low-income countries. International Journal of Educational Development, 31 (1), 2011, 3-14.

Rethinking Education from Birth: Building a Novel Learning Ecosystem for the AI Age
What if education did not begin at age three, five or six?
What if it began at birth?
And what if our goal was not simply to prepare a child for the next examination, the next grade or even the next job—but to develop, over the first twelve years of life, the foundations of a creative, intelligent, resilient, ethical and compassionate human being capable of continuing to learn throughout an unpredictable future?
This is the educational challenge of the Artificial Intelligence age.
We should not merely ask:
“How can we put AI into today's schools?”
We should ask something much more fundamental:
“If we were designing education today for a child who may live and work through the next 50 years of technological change, what would the entire learning ecosystem look like?”
The answer requires us to move beyond the idea of schooling itself.

From Schooling to a Birth-to-12 Learning Continuum
Traditional education largely begins when the child enters an institution.
A future-ready education ecosystem begins much earlier.
It recognizes birth to five as a foundational period in which relationships, language, sensory experiences, emotional security, play, curiosity and early executive capabilities are being formed. The accompanying Living Ecosystem framework therefore places the earliest years at the centre of the architecture rather than treating them as merely a prelude to "real education."
The ecosystem then evolves progressively:
Birth–2: Attachment & Sensory Foundation
Secure relationships, language, sensory exploration, movement, trust and emotional co-regulation.
2–5: The Critical Curiosity Years
Language explosion, imagination, guided play, self-regulation, storytelling, questioning and discovery.
5–8: The Bridge Years
Foundational literacy, numeracy, digital understanding, social learning, confidence and project-based exploration.
8–12: The Expansion Years
Interdisciplinary investigation, creativity, ethical reasoning, apprenticeship, collaboration, self-directed learning and emerging mastery.
The boundaries should be flexible.
The child should experience one continuous developmental journey, not a series of disconnected institutional transitions.

The Great Technology Journey Has Changed the Meaning of Education
Humanity has already experienced several information revolutions.
The Printing Press
Knowledge became reproducible and distributable at unprecedented scale.
It helped establish literacy and mass education.
Computers
Information processing became increasingly automated.
The Internet
Access to global knowledge became almost instantaneous.
Big Data
Patterns in human behaviour and learning became increasingly visible.
Machine Learning
Systems began identifying patterns and adapting to individual users.
Natural Language Processing
Machines became increasingly capable of interacting with humans through language.
Generative and Advanced AI
Machines became capable not merely of retrieving information, but of generating text, images, code, explanations, simulations and increasingly sophisticated forms of creative output.
Each technological revolution removed another information bottleneck.
So what remains uniquely important?
Human judgment.
Curiosity.
Original thought.
Creativity.
Empathy.
Ethical reasoning.
Collaboration.
Imagination.
Purpose.
The ability to decide what is worth doing.
This is why the AI age requires not less human education—but better human education.

The Parent Becomes the First Learning Architect
The first classroom is not a classroom.
It is the home.
The first teacher is not necessarily a professionally trained educator.
It is the parent or caregiver.
Parents therefore need to become intentional architects of the child's earliest learning environment.
Not by turning infancy into academic training.
Quite the opposite.
By creating a rich environment of:
Conversation.
Storytelling.
Play.
Movement.
Exploration.
Questions.
Affection.
Secure relationships.
Observation of nature.
Shared experiences.
Meaningful routines.
The framework emphasizes the power of everyday parent-child interaction, rich language, predictable rhythms, physical connection and responsive "serve-and-return" interaction during the earliest years.
During the curiosity years, parents should increasingly:
ask rather than merely tell,
read rather than merely instruct,
encourage rather than constantly correct,
allow productive struggle rather than immediately rescue,
and protect unstructured play and imagination.
AI can support parents by suggesting activities, providing developmental information, translating languages or reducing routine cognitive load.
But one principle must remain non-negotiable:
AI may support parenting. It must never replace the human relationship through which the child develops.

The Child Must Become an Active Agent
The child is not an empty vessel waiting to be filled with knowledge.
The child is a developing intelligence.
From the earliest possible stage, children should progressively experience:
Choice.
Exploration.
Experimentation.
Creation.
Failure.
Reflection.
Peer interaction.
Problem-solving.
Contribution.
Agency should therefore be deliberately developed.
A young child may choose between activities.
An older child may select a project.
A ten-year-old may investigate a question independently.
A twelve-year-old may design a personal learning pathway with a mentor.
The progression is:
Choice → Agency → Responsibility → Self-direction → Mastery
The framework explicitly treats children as co-authors of their learning through choice, productive struggle, peer teaching, reflection and increasingly sophisticated AI co-creation.

The School Must Be Reimagined as a Learning Hub
The future school should not disappear.
It should become more valuable.
But its purpose must change.
Instead of being primarily a centre for standardized content delivery, the school can become a Learning Hub that connects:
Home + Child + Peers + Teachers + Mentors + Community + Technology + Real-world Experience
Its principal functions should include:

1. Personalized Learning Pathways
   Every learner follows a developmental pathway combining foundational knowledge, interests, projects and emerging strengths.
2. Interdisciplinary Learning Studios
   Real problems bring disciplines together.
   A project on water may require:
   Science + Mathematics + Geography + Technology + Language + Design + Economics + Ethics.
3. Social-Emotional Development
   Empathy, communication, conflict resolution, resilience and ethical judgment become deliberate components of learning—not activities left over after academics.
4. Safe AI-Augmented Learning
   Children gain age-appropriate experience with AI through monitored environments, adaptive tutors, simulations and creative tools.
5. Community Apprenticeships
   Children increasingly encounter the real world through interactions with scientists, engineers, artists, entrepreneurs, farmers, craftspeople, healthcare professionals and community leaders.
   The accompanying model proposes precisely this evolution from conventional school toward flexible, community-anchored Learning Hubs with personalized pathways, interdisciplinary studios, safe AI environments and real-world apprenticeship.

Social Learning May Become More Important Than Ever
One of the greatest paradoxes of the AI age is that human-to-human learning may become more valuable precisely because machines are becoming more capable.
Children need to learn:
How to listen.
How to negotiate.
How to disagree.
How to persuade.
How to cooperate.
How to lead.
How to follow.
How to teach.
How to resolve conflict.
How to build trust.
These capabilities emerge through relationships.
Therefore, peer learning should not be an optional enrichment activity.
It should become a core architecture of learning.
A child who teaches another child strengthens their own understanding.
A child who collaborates with someone possessing different strengths learns the value of complementary intelligence.
Mixed-age learning communities can create powerful cycles in which younger learners observe older learners while older learners consolidate their own mastery by teaching.
The framework therefore places the learning pair and learning pod at the heart of the ecosystem.

AI Should Become a Learning Partner — Not a Learning Master
The future child should become fluent with AI.
But AI fluency must mean much more than knowing how to write prompts.
Children should learn to ask:
Why did AI produce this answer?
What evidence supports it?
What might be wrong?
What assumptions are hidden?
What perspectives are missing?
Can I verify this independently?
What can I create beyond what AI suggested?
This creates a new educational capability:
AI-augmented human intelligence.
The objective is not to make children dependent on intelligent machines.
It is to make them more intelligent humans because intelligent tools exist.
Adaptive AI tutors can provide personalized practice and immediate feedback.
Immersive technologies can enable experiences otherwise impossible, dangerous or expensive.
Simulations can allow children to explore complex systems.
But technology should be used selectively.
Not because it is impressive.
Because it improves learning.

Assessment Must Become a Continuous Portrait of the Learner
The traditional question is:
“What did the student score?”
The future question should be:
“What is this learner becoming capable of?”
Assessment should therefore continuously gather evidence of:
Knowledge
Reasoning
Creativity
Inquiry
Problem-solving
Communication
Collaboration
Practical capability
Resilience
Ethical judgment
Metacognition
Self-directed learning
Innovation
Instead of reducing the learner to one number, the system should build a living learner profile.
Projects.
Questions.
Experiments.
Designs.
Performances.
Peer feedback.
Mentor observations.
Self-reflections.
Academic assessments.
Community contributions.
The framework proposes a "talent constellation" rather than a single GPA, with AI helping mentors identify patterns while human educators interpret those patterns.

Every Child Has a Talent Constellation
Perhaps the greatest failure of conventional education is that it often asks:
“How good is this child at the things we have chosen to measure?”
A future-ready system should also ask:
“What extraordinary potential does this child possess that we have not yet discovered?”
One child may demonstrate extraordinary spatial intelligence.
Another may be a natural investigator.
Another may be a builder.
Another may possess unusual empathy.
Another may be a gifted communicator.
Another may be a future entrepreneur.
Another may see patterns others miss.
Another may possess exceptional artistic imagination.
The purpose of education should not be to make every child identical.
It should be to help every child discover:
Who am I?
What can I become?
What am I capable of creating?
What contribution can I make?

Well-Being Is Not an Add-On
A future-ready ecosystem cannot be built on exhausted children.
Play matters.
Sleep matters.
Movement matters.
Nature matters.
Nutrition matters.
Friendship matters.
Emotional security matters.
Mental well-being matters.
The framework deliberately protects unstructured play, outdoor activity, emotional literacy, movement and health as fundamental learning infrastructure rather than peripheral programmes.
A child should not have to sacrifice childhood in order to become academically successful.
Healthy development is the foundation of sustainable excellence.

Equity Must Be Designed Into the Architecture
A 50-year model cannot be designed only for affluent schools.
The future learning ecosystem must be scalable across:
urban and rural communities,
different languages,
different cultures,
different economic conditions,
different technological environments.
Technology can actually help democratize access if deliberately designed for equity.
Low-cost AI tutoring.
Offline learning resources.
Shared devices.
Community learning hubs.
Open educational resources.
Remote expert mentorship.
Local-language content.
Community educators.
The Hub-in-a-Box concept proposed in the framework offers one possible direction: modular, low-cost infrastructure capable of extending quality learning environments to resource-constrained communities.

Build for 50 Years — Not for the Next Technology Cycle
A genuinely future-ready education system should not be designed around today's AI tools.
Those tools will change.
The architecture must therefore separate:
What is permanent
from
What is technological.
Permanent foundations may include:
human relationships,
curiosity,
language,
reasoning,
creativity,
ethical judgment,
social learning,
physical development,
agency,
reflection,
continuous learning.
Technology layers can then evolve:
Printing → Computers → Internet → Big Data → Machine Learning → NLP → Generative AI → Whatever Comes Next.
The educational ecosystem should absorb technological change without repeatedly rebuilding itself.
That is the essence of a sustainable 50-year architecture.
The framework proposes exactly this principle: a modular system in which developmental stages and human roles remain stable while content and technology layers evolve, supported by periodic review and local adaptation.

The Ultimate Product of Education Is Not a Certificate
It is a human being.
Imagine a twelve-year-old emerging from such an ecosystem.
A young person who:
loves learning,
asks powerful questions,
can think independently,
is comfortable with uncertainty,
can collaborate naturally,
can communicate confidently,
has experienced failure without being defeated by it,
knows their emerging strengths,
can use AI without surrendering judgment,
can create rather than merely consume,
understands ethical responsibility,
cares about other people and the planet,
and most importantly—
believes they have the capacity to make a difference.
That is a very different definition of educational success.

A New Educational Compact
Perhaps the future education ecosystem can be expressed through one simple relationship:
Parents create the foundation.
Children exercise agency.
Peers create social intelligence.
Learning Hubs provide structure, mentorship and opportunity.
Communities provide authentic contexts.
AI amplifies human capability.
Assessment reveals growth.
The ecosystem continuously evolves.
When these elements work together, education stops being a sequence of institutional stages.
It becomes a living continuum of human development.

The Question We Must Ask Ourselves
The children born today may live into the next century.
Many of the occupations they will enter may not yet exist.
Many technologies they will use have not yet been invented.
Many problems they will confront are beyond our imagination.
So perhaps our responsibility is not to predict their future.
It is to develop the human capabilities that will allow them to create their own future.
The real transformation is therefore not:
AI in Education.
It is:
Education for the AI Age.
And that requires a much more radical shift:
From schooling to learning.
From instruction to development.
From subjects to capabilities.
From classrooms to ecosystems.
From marks to evidence of growth.
From passive students to active learners.
From isolated learning to social intelligence.
From technology consumption to human–AI co-creation.
From preparing children for known careers to developing people capable of creating the unknown.
The future of education will not be won by the institution with the most technology.
It will be won by the ecosystem that best develops human potential.
Our challenge is not to build smarter machines to educate children.
It is to build an education system capable of developing smarter, wiser, more creative, more resilient and more compassionate human beings who can live and thrive alongside those machines.
That is the education ecosystem we should begin building today—for the next 50 years.