Children need regular physical activity.
WHO recommends substantial activity for children and adolescents and supports whole-school approaches. Movement belongs inside the educational day without requiring every child to become an athlete.
Summary judgment
Chapter 11 remains one of the strongest Human Scale directions, but it becomes more plural. Education should develop knowledge, judgment, health, belonging, capability, and increasing independence without turning “unstructured,” “practical,” “traditional,” or “AI-powered” into automatic proof of quality.
What survives strongly
Sleep is an institutional design constraint.
CDC summarizes evidence connecting insufficient sleep with health and school-functioning problems and later school starts with more adolescent sleep in many settings.
School connectedness matters.
Students who feel cared for, supported, and like they belong show better health and educational outcomes on multiple measures. Belonging is valuable alongside academic quality, not instead of it.
What becomes more careful
Self-directed play is valuable without becoming a magic treatment.
Children need some genuine choice, exploration, imagination, movement, and peer negotiation. The amount and form should vary with age, disability, safety, culture, temperament, and setting.
Independence should mean growing judgment, not one nostalgic timeline.
Maturity, disability, traffic, crime, transportation, family life, culture, and place all change what safe independence looks like.
Practical skills matter, but school cannot teach everything.
Cooking, first aid, money, repair, digital security, civic navigation, and care can be useful. Families, schools, libraries, community organizations, employers, and vocational programs can share the work.
Multiple paths need real mobility between them.
University, community college, apprenticeships, vocational education, entrepreneurship, service, arts, certification, and portfolios can all work—if they create capability and do not become early traps.
AI in education
UNESCO now explicitly frames educational AI around a human-centered, rights-based approach, recognizing access and personalization benefits alongside privacy, safety, equity, governance, and dependency risks.
UNESCO GenAI Guidance ↗ · UNESCO Rights of Learners ↗
Human Scale test: after using the AI, what can the learner understand, judge, explain, or do that they could not do before—and what can they still do without it?
Measurement without metric worship
The answer to bad measurement is not no measurement. Use multiple measures where one proxy is too narrow; preserve comparability where it reveals inequity or failure; reduce administrative burden; and keep asking what human capability the metric is supposed to represent.
Recommended Chapter 11 changes
Movement + sleep: keep them strongly.
Belonging: measure it alongside learning, not instead of learning.
Play: protect self-direction without prescribing one amount or form.
Independence: measure growing judgment and competence.
Knowledge: keep serious foundational learning and method pluralism.
Practical skills: share responsibility across institutions.
Pathways: preserve movement between them.
AI: require capability-preserving design.
Measurement: keep it plural, meaningful, and lower-burden.
What would change our minds?
We should weaken self-directed-play claims if stronger evidence shows little benefit after key confounders or structured alternatives consistently outperform it without meaningful costs. We should weaken later-start advocacy where later starts fail to increase sleep or create larger local harms.
We should expand AI use where rigorous evaluations show durable learning, accessibility, teacher-capacity, and equity gains without replacing core skill formation or creating unacceptable privacy and dependency risks.