The fourth in The X-Challenges, a series of public challenges. P-20 means the whole ladder of schooling, from preschool through graduate school.
In brief. Governments are taking phones out of schools and writing rules for AI. Both may help, but neither reaches the cause, which lies in the subject-based curriculum and the classroom itself: one teacher facing one class, one subject at a time, an arrangement barely changed in four hundred years, from primary school to graduate school. The public pays for schooling, much of it with borrowed money, and families pay again in tuition and debt. No school or government ledger anywhere in the world counts the value of what children learn. The old bargain that a diploma buys a job is breaking down as AI arrives. Governments and billionaires have begun opening accounts for children, but money alone cannot teach a child to own, still less help a child grow into a life of their own.
Outside school, the internet children carry everywhere, all the time, was built by adults to sell their attention. Three X-Challenges came before this one. The first two were addressed to AI and to Elon Musk, but they were meant for everyone who built that internet and is now building AI: Google, Microsoft, Apple, Meta, Amazon, Alibaba, Tencent, ByteDance, JD.com, and the leaders of AI, among them Jensen Huang, Sam Altman, Dario Amodei, and DeepSeek’s Liang Wenfeng. The first asked for a clear, testable line between AI and human beings. The second asked for the next internet, built on what people own.
The third, Two Minutes, was put to everyone. No one can know what they will be thinking two minutes from now, because that thought does not yet exist. “The not-yet-known is not the absence of knowledge. It is the presence of possibility.” Yet “when education tells a child what to think before the child has formed the question,” the right to the not-yet-known “is quietly handed over.” So the internet and the school have to be rebuilt together. The challenges to AI and to Musk asked for a new internet. This one asks for a new school, where children learn and discover something worth owning. In the age of AI, neither will hold without the other.
This X-Challenge rests on one idea: what a learner discovers, makes, and comes to understand is a valuable asset, and it belongs to the learner. From that idea come four challenges to every school and every government. Build the infrastructure for authentic learning, and let every learner own their learning outcomes. On it, redefine learning, the teacher, and the school, and rebuild the school for the future. In that school, let every child and young person grow into what they are capable of, reaching the cognition of those now twice their age. From there, open a new Age of Cognitive Exploration and Innovation for all of humanity. All four begin with children, with ownership, with the local, and with life as it is truly lived.
No one can build this infrastructure alone. It is a network, like the phone, which is worth having because the people who matter to you have one too. Steve Jobs could not have built a phone for his own children alone; Mark Zuckerberg cannot build this network for his, nor can Harvard or Stanford build it for their students. It can start in one community, but it must be built so that every child can join and connect.
Nor is it a utopia. Since 2012, a public school in Beijing, starting from its architecture and its organization of learning, has redefined learning as authentic learning and, on the basis of learning relationships, rebuilt its learning spaces and the school itself. For ten years, Ann Lieberman and Ray Pecheone of Stanford, together with leaders of innovative schools in Finland, Israel, Australia, the United States, and elsewhere, and their schools, students, and teachers, walked beside it.
The wrong tree
Phones are now kept out of classrooms in more than half the world’s countries. By March 2026, UNESCO’s Global Education Monitoring Report counted 114 education systems, in 58 percent of the world’s countries, that restricted them by national law or policy, up from 24 percent in June 2023. In the European Union, 17 of the 27 member states ban phones for the whole school day or in class. In many schools the day now starts at a pouch or a locker, where the phones go in before the children go to class.
The bans help. They are meant partly to protect children’s mental health, and the concern is real. In 2023, nearly 40 percent of American high school students reported persistent sadness or hopelessness. A ban can remove a distraction. It cannot give a child a reason to learn.
As phones go out of classrooms, AI comes in. In the 2026 legislative season, FutureEd, an education policy think tank at Georgetown University, counted 87 bills on AI in the classroom across 28 American states; 13 have become law. Most set guardrails: which tools, which data, when parents may opt out.
Both moves treat the technology itself as the danger, and so bark up the wrong tree. What matters more is whom the technology was built for and who owns it. The internet in children’s pockets was built for the adults of industrial society and was offered to them as a commons. For most people it never was one. We share the use; the data goes to the platforms. The business runs on attention, which means it runs on us, and adults pay for it as surely as children do.
A phone ban shields children from that system for a few hours a day. At eighteen they enter it anyway, and for all they learned and did in school, they carry out no evidence of their own: nothing they control, and can keep adding to and using.
Fencing off this attention-based internet, or patching it, will not be enough now that AI has arrived. Since ChatGPT was released in November 2022, AI has come to carry much of what humanity has written down over thousands of years, and it can sit beside every learner, one at a time. For the first time, learning need not be organized around one teacher and one class.
Whom will it serve? Built like the old internet, free to use and paid for with attention, AI will become the old internet at closer range, holding not only our attention but the way we think. Built on learning ownership, it will let every learner keep their learning outcomes for life, along with the evidence of what they studied, what they learned, and what they came to understand. The old internet’s business model has already followed AI in: in early 2026, ChatGPT began testing ads, which OpenAI says are kept apart from its answers.
Whom AI will serve is the question that runs through every X-Challenge, and it is why the school and the internet have to change together. A new internet and a new AI with no authentic cognitive outcomes to carry will drift back to selling attention. A new school whose children’s learning outcomes are still controlled by someone else’s platform is handing them away. The infrastructure this fourth X-Challenge asks for is where the two meet.
The Mismatch
The problem is older than phones or AI. The classroom has barely changed since the Czech educator Comenius set out class teaching as a system in 1632: children of the same age, one teacher, a bell, the next subject.
The X-Challenges calls the result the Mismatch, an industrial shell around an agricultural core. The shell is the whole industrial apparatus built around teaching and learning: standardized curricula, fixed timetables, grade after grade moving down an assembly line, accountability by testing, government education funding, school management, financial accounting, and the ratio of teachers to students. The core is teaching and learning itself, and its tools are those of the agrarian age, before machines. One teacher stands before a room of children with a blackboard, a voice, two hands, and a pair of eyes.
So long as the core stays as it is, technology, AI included, has nowhere to work but the shell, where it adds more tests, more data, more forms. The harder technology works on the shell, the more it squeezes the teachers and children inside, and many of them break under the strain, in burnout and in mental distress.
In America, farms have long since been mechanized. The classroom has to change too, but not the way farms did. What grows there is children, ours and everyone’s.
Two pillars hold the shell to the core: class-based organization and subject-based instruction.
The classroom
The core, one teacher facing one class, is what the X-Challenges calls the classroom, whatever its size. Take a class that is small by American standards: eighteen students and a teacher, nineteen people in all. Count each person’s influence on every other person as a separate thread, A on B and B on A, each either present or absent. The possible patterns of learning relationships then number 2^342 (2 to the 342nd power), a figure 103 digits long, more than the number of atoms in the observable universe. In a class of forty, common in China and across much of Africa, the figure runs to 494 digits. No teacher could manage every one of those threads, and none should be asked to. The trouble is that a classroom resting on one adult cannot see most of the relationships through which children actually learn, much less nurture them.
Teachers know this. It is why classroom management and discipline come first in teacher education, and in schools from preschool to graduate school. When the relationships cannot be held, the children are held instead. The classroom is meant to teach respect, but its structure keeps pushing teachers toward maintaining obedience.
The university is a bigger version of the same classroom, and there the second pillar, subject-based instruction, is in plain sight. Departments and majors divide knowledge, credit hours measure time in a seat, one subject at a time, and a lecture hall turns hundreds of faces toward one.
Technology has promised to change the classroom for more than a century. In 1913, Thomas Edison predicted that “books will soon be obsolete in the public schools.” In 1996, Steve Jobs, who by his own account had given more computers to schools than anyone, concluded: “What’s wrong with education cannot be fixed with technology.”
Money has tried too. In the 2000s, Bill Gates’s foundation spent hundreds of millions of dollars remaking American high schools, and in 2009 he wrote: “Many of the small schools that we invested in did not improve students’ achievement in any significant way.” A later $575 million effort to raise teacher effectiveness did not substantially improve student outcomes, the RAND Corporation found in 2018. Mark Zuckerberg’s $100 million for Newark’s schools, announced in 2010, was meant to build a reform model that could spread across the nation. That ambition came to nothing.
The world spends roughly $7 trillion a year on education and training, and more than a billion children are in school. At that scale, even the largest fortunes and the richest companies on earth are small. From outside, neither technology nor money can change the system. It has to be reformed from within. Yet the trillions spent within it each year have not changed the classroom either; with technology’s help, they have reinforced the Mismatch.
In 2023 Khan Academy launched an AI tutor, and in 2025 Google opened its AI to students. If the classroom stays as it is, there is little reason to expect these tools to do what donated computers and philanthropic millions could not. Phone bans, AI rules, computers, and money all worked on the shell and left the core untouched, because no one can change what they cannot see.
Who pays
The public pays for schooling twice. Governments raise taxes and pay for seats, staff, and buildings, and what they get back is measured mainly in grades, admission rates, and diplomas, which show years served and tests passed but say little about what a child has learned and can do. Families then pay a second time, in tuition and other related costs.
An American school district keeps its accounts by object of spending: salaries, benefits, purchased services, supplies, property, debt. Not one line lists or notes how much real learning any of it bought.
Every school budget is drawn up in the children’s name. Ask what the children have learned, and the system can still answer only with the same scores and diplomas.
That answer was enough while a diploma led to a job. It no longer reliably does. In the United States in 2026, recent college graduates are more likely to be unemployed than workers as a whole, and in the jobs most exposed to AI, employment of 22- to 25-year-olds has fallen about 19 percent relative to their peers. In a Gallup survey released in September 2026, only 32 percent of Americans said they were satisfied with the quality of K–12 (primary and secondary) education, the lowest share in the 27 years Gallup has asked.
Governments invested, teachers taught, students learned and made things. Yet no one, and no government, can tell students and their families: What did they learn? How does anyone know? Having learned it, what then? And if they did not, what then?
That is a black hole in family budgets, in government finance, in schooling, and in the national economy. Grades and diplomas never filled it, and now that AI can write the essay and pass the test, they can no longer even hide it.
Governments know what they spend. What they lack is a system that links public investment to real outcomes that learners own and can keep using. A ledger with no line for the value of learning cannot tell an investment from a loss.
Much of this money is borrowed, directly or indirectly, by governments whose books count education only as a cost. America’s federal debt has topped $40 trillion, and by late 2025 the government was spending more on interest than on defense. The pensions promised to teachers are short about $562 billion in the state funds alone. In China, local government debt reached about 54.8 trillion yuan (roughly $7.7 trillion) in 2025, more than the central government’s own.
Schools did not cause these debts. But under the industrial model their budgets keep rising, and the books show only what went out. National accounts count education as consumption rather than investment, so whenever debt bites, education looks like a cost to cut. The World Bank estimates that human capital makes up 64 percent of the world’s wealth, yet it appears on no nation’s balance sheet. A country that cannot see its largest asset cannot invest in it wisely, and a government that borrows against the future without counting the value of what its children learn has no way to judge what the borrowing buys. The way schools are organized, funded, and measured has become economically and socially unsustainable.
The share that families pay eventually lands on millions of students themselves. They spend sixteen years or more in school, borrow to get through it, and leave with little to show for it but a diploma and a debt. American student debt nearly tripled between 2008 and 2023, to $1.77 trillion. By mid-2026, more than 9.3 million federal borrowers were in default, owing about $234 billion. Default carries the threat of garnished wages and seized tax refunds, and in October 2026 the Treasury was reported to be preparing to hire private debt collectors to pursue those borrowers, possibly as soon as November.
A debt signed at eighteen can follow a graduate past sixty. What families and students paid, or borrowed, rested on one promise: that a diploma leads to a job. That is the promise now breaking down. Graduates leave owning the debt but not the evidence of what they learned.
Three Nobel Prizes and the changing education bargain
That promise has been explained by economics that won three Nobel Prizes. Theodore Schultz and Gary Becker, laureates in 1979 and 1992, showed through human capital that schooling is an investment in people, because knowledge and skill raise what a person can produce and earn. Michael Spence (2001) showed through signaling why a diploma pays: employers cannot see ability directly, so they read the diploma as a signal.
These theories explain a bargain that already existed between education and the economy, and they may well still hold. What is giving way is how the industrial school has kept that bargain.
Schultz put it plainly in 1961: “Laborers have become capitalists … from the acquisition of knowledge and skill that have economic value.” The capital was always in the person. The proof of it never was.
The industrial school kept the bargain on two conditions. It made itself the middleman of knowledge, standing between children and what humanity knows and deciding who passed, and it made grades and diplomas the proof of what had been learned. AI has removed the first condition and begun to remove the second. What the middleman once guarded is now a question away. What happens to the second, one American university has already found out.
In May 2023, at Texas A&M University–Commerce, an instructor pasted his students’ essays into ChatGPT, asked it whether it had written them, and gave much of the class an incomplete; some diplomas were put on hold. Some students defended themselves with the edit histories of their Google Docs, and the university later said no one had failed or been barred from graduating.
The whole P-20 system of learning knowledge is giving way, and the logic of the ladder from preschool to graduate school is failing with it. For two centuries the logic was simple. Fix the curriculum and the majors, and the jobs would follow. A country that fell behind could catch up by following the road the leaders had taken. AI has undercut both bets.
When a major no longer leads reliably to a job, what links learning to life has to be something else: the learner’s own authentic learning outcomes, achieved on real problems, with real people, in the learner’s own family and community. Helping every learner build learning outcomes of their own is what this X-Challenge asks of every school and every government.
The four challenges
The X-Challenges first put these four challenges to the world in November 2023, above all to governments and institutions. Here they go to the schools themselves, and to the governments behind them.
First, build the infrastructure for authentic learning, and let every learner own their learning outcomes. This value-based infrastructure is digital, Locobal, decentralized, and generative, powered by AI; it connects with teaching and learning and offers authentic, innovative learning experiences. Locobal, an X-Challenges word for local self-directed globalization, means rooted in each place and connected with other places: not a globalization that dissolves the local, but a global connection built on strengthening it. The learning outcomes held there belong to the learner and are protected by law.
Second, on that infrastructure, redefine learning, the teacher, and the school, and rebuild the school. Schooling moves from manual labor to human-led learning, powered by AI. Learning becomes discovery, study, research, and innovation, wherever learners are and whenever meaningful relationships form. Teachers become the architects of learning relationships and the professional generators of the future assets and values of families and communities. The school becomes the creator and hub of a learning ecosystem, with a mission to educate children, unite people, and lead society.
Third, in that new school, let every child and young person grow into what they are capable of, reaching the cognition of those now twice their age. Every child, not a chosen few, should have the chance to develop new capabilities and to lead in a time of great change. Today, when a teenager finds what professional scientists have missed, it is often luck, or birth. Schools should cultivate such discoveries for every child rather than leave them to luck and birth.
Fourth, from there, open a new Age of Cognitive Exploration and Innovation for all of humanity, one that develops the untapped capabilities of every child and opens a way to reshape the future of human civilization. Each generation can explore what lies close at hand, the countryside, the grassland, the farm, the coast, the city block, and bring what it finds into exchange with the world. The first Age of Exploration set out from a few nations and found new continents. This Age of Cognitive Exploration belongs to all of humanity. It sets out from wherever a child lives, and what it seeks is the untapped capability in every person, and continents of cognition not yet perceived.
All four start in the same places: with children, with ownership, with the local, and with life as it is truly lived.
A government takes up this challenge with three acts. It recognizes in law each learner’s right to their own learning outcomes. It builds the infrastructure in at least one school and community and, with the learners’ consent (or, while they are children, their families’), shows the public what they made and what it brought. And it adds the missing line to its education accounts: children’s learning outcomes and their value, meaning what that learning brings to the children themselves, their families, and their communities, drawn from learning outcomes the learners own rather than from another standardized score. That is the line by which the return on public money should be judged. The three acts are only a start.
A school takes up this challenge by rebuilding itself on that infrastructure, so that every learner leaves with a range of authentic learning outcomes of their own. The challenge succeeds when every learner in a given place holds the right to their own learning outcomes for life.
Learning ownership
By learning outcomes, the X-Challenges means what a learner discovers, makes, and comes to understand, not test scores. Learning ownership means recognizing those outcomes as valuable assets that belong to the learner, held in trust by the family while the learner is a child.
Children can inherit land and hold shares, and now they can own what they know. Yet today the evidence of their work, growth, and understanding is held by others, and used without their say.
Some will say children already own their homework and their work, and in a narrow legal sense they do: an essay a child writes is the child’s, and they may keep every page. But at the end of the school year much of it is thrown away, and what survives is a grade, kept by the school. A study trip, a summer abroad, a project for a neighbor, a question pursued for a year rarely leave evidence behind: what was learned, how anyone knows, how well, and what if it was not. To own learning outcomes is to be able to keep them, use them, build on them, and have their origin known when others use them.
Handing them to the school for safekeeping is no guarantee either. In December 2024, a hacker broke into PowerSchool, a company whose software serves more than 60 million students. In some districts the stolen data included grades, medical conditions, disability accommodations, and free-lunch status, a childhood’s worth of information held by a company many families had never heard of.
A child’s learning outcomes can be of value to the child, the family, and the community, and that should be their most fundamental and greatest value. No town on earth can list every plant and insect within its limits; there are not enough scientists to count them. Children can help. In 2019, two Bay Area high school students, Harper Forbes and Prakrit Jain, were combing through iNaturalist, a free app where people share photos of living things and identify them together, when they spotted a scorpion no one could name. With Lauren Esposito of the California Academy of Sciences, they went on to describe two species new to science, published in 2022 in the journal ZooKeys. One of them, Forbes warned, could be wiped out by a single solar farm, mine, or housing development. Their work had scientific value, public value for decisions about land, and personal value, since their names are attached to a discovery. Children who survey the life of their own street learn language, mathematics, science, technology, and engineering along the way, and the accounting and finance that come with it; they also learn how to treat themselves and others, and how to work with them.
Yet children still sit in classrooms with the words and numbers of a textbook, waiting to be measured by exams and by international assessments like PISA. Those days should not last long. The living world outside the door waits to be studied and discovered, by each child from where they stand. Doing that, with real scientists, on things that truly matter, is authentic learning. What is missing is a tool that lets both the process and the outcomes of every child’s exploration be kept and used, and proves they are the child’s own.
In 2005 Britain began opening a trust fund for every child born from September 2002 to January 2011. Since 2001, Singapore has matched, dollar for dollar up to a cap, what parents save in each child’s development account. In 2026 the U.S. Treasury began seeding an investment account for every eligible child born from 2025 through 2028, with $1,000 placed in an index fund and locked until the child turns eighteen. In December 2025, Michael and Susan Dell pledged $6.25 billion to seed similar accounts for some 25 million children aged ten and under, and in June 2026 Micron pledged $250 million to the accounts of up to a million children. On October 19, 2026, Elon Musk pledged to give one of his SpaceX shares to every child in America next year.
Governments and some of the largest fortunes on earth now accept that a child can own capital from birth and let it compound. That money comes from outside. What children learn, they earn themselves, and it compounds for life, yet it has no account of its own.
Owning money is not the same as knowing how to own. Britain’s trust funds have begun to mature, and by September 2025, 758,000 matured accounts belonging to young people aged 18 to 23 had gone unclaimed, worth about £2,240 (roughly $3,000) each on average. An account young people do not know they have, or do not know what to do with, is theirs only on paper.
Nor is this only a matter of the young. In a 2014 survey of more than 150,000 adults in over 140 countries, only one in three understood at least three of four basic ideas: interest, compounding, inflation, and spreading risk. In the United States, the share was 57 percent. The P-20 ladder seldom teaches what it takes to own: how money grows, what a debt costs, how to read a balance sheet. A society that cannot read a balance sheet will also find it hard to help its government work out what the borrowed money actually bought.
Give children a fish and you feed them for a day; teach them to fish and they can feed themselves for life, and pass the skill on to the generations after them. The accounts hand out fish. Teaching children to learn and to own is teaching them to fish, and they learn to own by owning what they learn, from the start.
Today, whether a child’s learning outcomes are ever seen depends largely on luck, and luck follows rank: whose school, whose family, whose name, how many followers. The sociologist Robert Merton named this the Matthew effect in 1968, after the verse “unto every one that hath shall be given.” Online it has hardened into a cognitive feudal order, in which attention is inherited and a follower count works like a title. Social media did not invent that order, but it magnifies it, and with every scroll children take it in as simply the way the world works.
Children whose learning outcomes are their own do not have to wait for someone higher up to notice them. Every child should have the right to show what they have learned and made, prove it is theirs, and be credited for it.
No ownership, no democracy. A society that opens accounts for its children and lets them inherit property, while denying them ownership of their own learning outcomes, cannot go on as it is.
Learning ownership needs three things.
A right. What learners make, and the evidence of what they learn and come to understand, belong to them, at every age from preschool through graduate school. For children, the family holds these in trust and hands them over as the child grows.
An architecture. These assets are held by the learner, not by the school, the platform, or the AI company, and they remain the learner’s for life.
A rule of use. No school, platform, or AI company may sell them, and none may profit from them or train AI on them without the learner’s consent, or, for a child, without the family’s consent given in the child’s interest.
In law, none of this is yet one thing. Rights in a child’s work, personal data, proof of what a person can do, and knowledge itself fall under separate rules, and few of those rules let children act on their own behalf. Even where children hold a right, they lack the tools to use it: where to keep what they learn, how to prove it is theirs, who may use it, and who decides.
AI provides new conditions for such tools to reach every learner. Learners could keep their learning outcomes, and the proof of what they can do, as they learn and make, instead of waiting for a single diploma to sum it all up at the end. Sorting out the law is work for lawmakers, and part of this challenge. Building the tools is the first of the four challenges.
The teacher
Lisa Wolfe, a middle school teacher, taught for nineteen years before she quit in October 2023, calling the job “soul crushing.” “Education only works off of free labor from teachers,” she told Fortune. In a RAND survey in early 2024, American teachers reported working 53 hours a week, nine more than comparable professionals, for $18,000 less in base pay.
In The Mismatch, the X-Challenges describes teachers’ earnings as a “capped limited income.” But the same system caps more than their income: it also limits students’ potential and the return on educational investment.
Nor is it only a matter of pay. Even in Finland, known for the respect it gives teachers and for requiring them to hold a master’s degree, teachers have gone to the streets again and again. In September 2015, members of the teachers’ union OAJ joined a national protest against government spending cuts, and schools across the country closed. In spring 2022, Finnish teachers struck for better working conditions for the first time since 1984: more than 22,000 educators in ten cities, among them Helsinki, Espoo, Turku, and Tampere, and for a week in May many schools shut. They won a pay programme of about 5.1 percent above general raises, to run until 2027; in 2025, the local government employers terminated it. In April 2026, teachers at Finland’s universities of applied sciences struck, under OAJ, for the first time; the strikes came in waves and spread to several institutions before a new agreement was reached in September. In August 2026, an OAJ survey found that 14 percent of teachers had faced violence at work in the past year, up from 10 percent a year earlier, and the union warned that the workload was making the profession ever harder to recruit into and stay in.
The shell asks teachers to deliver standardized content, grade it, fill in the forms, and keep order in a room no one can hold. Little time is left for what they do best: knowing each child, building the relationships in which learning happens, helping a young person find something worth doing.
AI changes the division of labor. It can take on the content, the drills, and much of the paperwork, and it can help with grading. It cannot know why a child is quiet today, nor how that child should choose what to do. That stays with the teacher.
Whether teachers can do it depends on redefining the teacher: as the architect of learning relationships, sharing responsibility for each child with colleagues, and helping each learner build up authentic learning outcomes of their own. What learners make belongs to the learners. The teachers who help them make it are the professional generators of those assets, and their part should be seen and rewarded.
The core of the classroom has always been the teacher’s. Teachers should lead its rebuilding, as they have at the first 3.0 Future School since 2012. They need new tools and a new system that support their work and their care for children.
The pioneering practice
In 2012, as a project approved by the Beijing municipal and Haidian district governments, Zhongguancun No. 3 Primary School, a public school in Beijing, set out on its journey as the first 3.0 Future School. Starting from its architectural design and from the way the school organizes learning, it began to rebuild class-based organization and subject-based instruction. It rests on four cornerstones: authentic learning; learning relationships; the neighbors’ kids; and the family–school community.
The neighbors’ kids means children of different ages learning from one another and walking side by side, as neighbors’ children do, while adults encourage and support every child as they would a neighbor’s. One class from each of three grades formed a cluster that spent every day together under one team of teachers, drawn from different subjects and sharing responsibility for each child.
Since 2016, every sixth-grader graduating from the school has completed an integrated capstone project and defended it before parents, the community, and the public. From 2017 until the pandemic, China Education Daily live-streamed the defenses online.
For ten years, Ann Lieberman of Stanford’s Center for Opportunity Policy in Education, Ray Pecheone of its Center for Assessment, Learning, and Equity, and Faye Snodgress of Kappa Delta Pi, the international honor society in education, walked beside the school, joined by leaders of innovative schools in Finland, Israel, Australia, China, the United States, and elsewhere, and by their schools, students, and teachers. More than 100,000 students and teachers have experienced authentic learning together.
Then the school ran into a wall that no single school could get past. Its children carried out authentic learning and defended it in public, yet the ladder ahead of them, from admissions to the job market, still counted only grades and diplomas. There was no infrastructure to hold their learning outcomes, and no law said those outcomes were theirs.
No school, Harvard and Stanford included, can bring that wall down alone. That is why the infrastructure comes first among the four challenges, and why it has to be built for every child.
Why now
In 2023, the X-Challenges warned that any nation that did not, within five to ten years, secure learning outcomes in law as valuable assets would lose its edge. Three of those years are gone, and at most seven remain. The first community or country to secure every learner’s lifelong ownership of their learning outcomes may take the lead in the age of AI, and in the future it shapes.
AI moves fast. The time to begin is now.
Human beings created AI. They should not have to live under it, or in fear of it. Begin with children, with ownership, with the local, and with life as it is truly lived.
The disruption has begun, and no one will escape it entirely. What is still open is how young people meet it: graded and judged by others, or owning their learning and their learning outcomes, with the right to decide how both are used. Schools will remain, as they should; what has to change is what they are built on. Teachers should lead that change, as they have always been expected to.
That day should not be far off: when children walk out of school, every one of them owns learning outcomes of their own. That is who they are.
From The X-Challenges, in seven volumes: The Mismatch (Volume Two) and 3.0 Future School (Volume Three).
Notes, by section
The wrong tree. UNESCO Global Education Monitoring Report, country monitoring of phone restrictions in school, as of March 2026; the June 2023 figure is from the 2023 GEM Report, Technology in Education. ENESET, analytical report for the European Commission, Mobile Phone Bans in Schools across the EU, 2026: by its country review, 17 of 27 member states have full-day or in-class bans. U.S. Centers for Disease Control and Prevention, Youth Risk Behavior Survey 2023 (MMWR Supplement, 2024): 39.7 percent of high school students reported persistent sadness or hopelessness. FutureEd, 2026 State AI in Education Bills tracker, updated October 6, 2026. OpenAI, ad testing in ChatGPT, announced January 2026.
The Mismatch. The X-Challenges, The Mismatch: Global P-20 Education’s Industrial Shell and Agricultural Core. John Amos Comenius, The Great Didactic (written in Czech in the 1630s; Latin edition 1657).
The classroom. Learning relationships: directed model, 2^[n(n−1)]; n = 19 gives 2^342 ≈ 8.96 × 10^102 (103 digits); n = 41 gives 2^1640 (494 digits). Average U.S. public school class size, 2020–21: 19.1 in self-contained elementary classes, 21.0 in departmentalized secondary classes (NCES, National Teacher and Principal Survey). China’s class-size standards cap primary classes at 45 and junior secondary at 50; classes of about forty are typical. UNESCO Institute for Statistics, School Resources and Learning Environment in Africa (2016): most single-grade primary classes in the countries surveyed have 40 or more pupils. Thomas Edison, interview, The New York Dramatic Mirror, July 9, 1913. Steve Jobs, interview with Gary Wolf, Wired, February 1996. Bill Gates, 2009 Annual Letter, Bill & Melinda Gates Foundation. RAND Corporation, evaluation of the Intensive Partnerships for Effective Teaching, 2018. Mark Zuckerberg’s gift to Newark Public Schools, announced September 2010; see Dale Russakoff, The Prize (2015), and Russakoff’s interview with NPR, September 2015: the aim was to build, within five years, a model of reform that could spread to other cities. HolonIQ, global education and training expenditure, estimated at $7.3 trillion for 2025. Google, Gemini for Education, opened to students in 2025. Khan Academy, Khanmigo AI tutor, launched March 2023.
Who pays. National Center for Education Statistics, Financial Accounting for Local and State School Systems, 2014 edition (expenditure object codes). Federal Reserve Bank of New York, The Labor Market for Recent College Graduates (ages 22–27), 2026. Stanford Digital Economy Lab, “Canaries in the Coal Mine?”, August 2026 update. Gallup and Walton Family Foundation, survey of U.S. adults, conducted August 2026, released September 2026. U.S. federal debt above $40 trillion, 2026 (AP); net interest exceeded defense spending in October–December 2025 (U.S. Treasury data). Equable Institute, 2025: teacher pension funds managed by the 50 states and D.C. short $562.2 billion. China local government debt about 54.82 trillion yuan and central government debt about 41.23 trillion yuan, end of 2025 (Xinhua, via Yicai Global). World Bank, The Changing Wealth of Nations 2021: human capital 64 percent of global wealth in 2018. Federal Reserve, student loans outstanding, Q1 2008 ($619 billion) and Q1 2023 ($1.77 trillion). Federal Student Aid, quarterly portfolio data released September 22, 2026: more than 9.3 million borrowers in default as of June 30, 2026, about $234 billion. The Washington Sun, October 19, 2026, and Forbes, October 19, 2026: Treasury preparing to hire private debt collectors, possibly as soon as November (Treasury and OMB declined to comment). U.S. Department of Education, delay of involuntary collections (wage garnishment and Treasury offset), January 16, 2026; defaulted loans remain in default.
Three Nobel Prizes and the changing education bargain. Nobel Prize in Economic Sciences: Theodore W. Schultz (1979), Gary S. Becker (1992), A. Michael Spence (2001). Theodore W. Schultz, “Investment in Human Capital,” American Economic Review 51, no. 1 (1961). Texas A&M University–Commerce, May 2023: instructor’s use of ChatGPT to judge student essays; university statement that no student failed or was barred from graduating, as reported by PCMag and others.
The four challenges. xchallenges.com, November 2023. Locobal (local self-directed globalization) is an X-Challenges term, from its LLG framework: Local, Locobal, Global. Definitions of learning, the teacher, and the school: xchallenges.com, “Our Story.”
Learning ownership. Jain, Forbes, and Esposito, “Two new alkali-sink specialist species of Paruroctonus,” ZooKeys 1117 (2022); California Academy of Sciences release, August 2022: the first unidentified scorpion was spotted in 2019. PowerSchool breach, December 2024: PowerSchool states its software serves more than 60 million students; data types reported by affected districts (TechCrunch, January 28, 2025). S&P Global FinLit Survey (Klapper, Lusardi, and van Oudheusden, Global Financial Literacy Excellence Center), released November 2015, fieldwork 2014; U.S. adults financially literate: 57 percent. United Kingdom, Child Trust Fund, launched 2005 for children born from September 2002 (closed to new accounts 2011); HM Revenue & Customs, press release, September 30, 2025: 758,000 matured accounts of 18- to 23-year-olds unclaimed, average value about £2,240. Singapore, Child Development Account (Baby Bonus scheme, since 2001), government co-matching of parents’ savings. U.S. Department of the Treasury, press release on Trump Accounts, January 28, 2026; accounts launched July 2026. Michael and Susan Dell pledge, December 2, 2025, for children aged ten and under. Micron Technology pledge of $250 million, June 30, 2026. Elon Musk, post on X, October 19, 2026, as reported by Axios and TechCrunch; a pledge, not yet carried out. Robert K. Merton, “The Matthew Effect in Science,” Science, 1968; the name comes from Matthew 25:29.
The teacher. Lisa Wolfe, Fortune, December 17, 2023. RAND Corporation, State of the American Teacher survey, January–February 2024. Finland: Yle, September 13, 2015: OAJ teachers to join the national protest of September 18, schools closed nationwide. OAJ and Juko press release, March 21, 2022: strike warning covering more than 22,000 OAJ members in ten cities; SOOL (Finnish student teachers’ union), May 5, 2022: teachers striking for better working conditions for the first time since 1984; JHL, May 4, 2022: strikes of May 3–9 in ten cities, schools closed. Kaleva, April 28, 2025: the 5.1 percent pay programme agreed in 2022 to run to 2027 terminated by the local and wellbeing-services employers (KT). Yle, April 8, 2026: first strike by teachers at universities of applied sciences, beginning at Metropolia, with strike warnings extended to six more institutions; Yle, September 2026: strikes at six universities of applied sciences, September 1–3; OAJ press release, September 11, 2026: OAJ accepts the mediator’s proposal and the new agreement takes effect. Yle, August 5, 2026: OAJ survey, 14 percent of teachers faced violence at work in the past year (10 percent in 2025).
The pioneering practice. Stanford SCALE and the reach of the 3.0 Future School: xchallenges.com, “Our Story.” Ministry of Education of the People’s Republic of China, “Beijing Zhongguancun No. 3 Primary School Explores the Practice of Systemic School Change” (in Chinese), April 23, 2018. Graduation defenses live-streamed by China Education Daily, 2017–2019.
Why now. The five-to-ten-year warning: xchallenges.com, November 2023. The X-Challenge to AI, published on X, September 2026. The X-Challenge to Elon Musk, published on X, 2026. An X-Challenge to Everyone: Two Minutes, published on X and WeChat, October 2026.
Cases, comments and nominations: research@xchallenges.com
All public challenges: Public Challenges