We are preparing children for futures we cannot imagine while rewarding them for being predictable.
It is a peculiar way to approach the next decade. We ask schools to cultivate initiative, then organize much of their work around compliance. More capable technology gives us new ways to enforce compliance, too.
The contradiction ran through my conversation with Katherine Prince at EdWeek Market Brief’s K-12 Market Futurecast on September 15. We were asked what school districts would need in 2036. Again and again, we came back to what schools allow people to do now: how much judgment teachers can exercise, how much responsibility students can assume, and how much influence either has over the systems shaping their education.
Michelle Caffrey, Market Brief’s Assistant Managing Editor, hosted the discussion as part of the publication’s tenth-anniversary programming. I appreciated her willingness to take a conversation for education companies well beyond the next purchasing cycle. Katherine, Vice President of Foresight and Strategy at KnowledgeWorks, brought insights from Charting a New Course for Education, particularly its exploration of changing educational structures and the future of education as a public good (Crabtree et al., 2025).
A question connects this conversation with my work on preparation, educational sovereignty, and positive rebellion: how can schools develop people’s capacity to shape the future if teachers and learners have so little influence over schooling itself?
Institutional imagination has fallen behind
Ten years ago, we were talking about tablets, STEM, STEAM, and the conviction that every child needed to learn to code. AI has since complicated the promise that coding would provide a dependable route into employment. Yet the classroom remains recognizable across a much longer span of time. Children occupy age boxes, knowledge occupies subject boxes, and the day occupies time boxes. We have become very good at putting screens in the boxes.
The familiar structure can conceal real changes within it. As Katherine observed, personalization and learner-centered approaches have advanced, and connections between schools and other learning environments have grown. The broader shift toward learning anytime and anywhere has been less extensive than many anticipated. A school can incorporate considerable technical innovation while leaving its expectations of learners almost untouched.
In Unprepared: Education beyond 2030, I question how far education can rely on anticipating future requirements and preparing learners to meet them (Moravec, 2026c). Developing a curriculum takes time. So does educating a person. The work, technologies, and institutions a program prepares students to enter can change before they graduate. Knowledge remains valuable; promises about where a qualification will take you are harder to defend.
An audience member asked a fair question: when was the future ever certain? Uncertainty alone cannot explain the problem. My concern is that technological, economic, and environmental changes can overtake the assumptions on which educational plans depend, while schools remain bound to those plans. Telling graduates to keep adapting gives them responsibility for a problem they have little power to address. Institutions need the capacity to reconsider their commitments, too.
Our difficulty imagining different schools compounds the problem. Even in science fiction, schools of the future often retain the familiar classroom. Apparently, imagining extraordinary machines is easier than imagining a different relationship between children and adults.
Who makes the educational decisions?
When schools depend on vendor-controlled AI, responsibility can become separated from authority. Teachers answer to students and families for uses of systems whose underlying models, data practices, and updates they do not control. Michelle put the question directly to us: are we asking schools to manage too much of the risk while vendors retain control?
Consider the decisions already embedded in an AI learning product. Which sources does it treat as authoritative? What does it accept as a good answer? When does it offer assistance, and when does it allow a learner to struggle? These choices affect education before a teacher writes a lesson plan. As a product becomes indispensable, its assumptions can become the school’s habits.
My paper Beyond the Digital Enclosure: AI, Coloniality, and the Pursuit of Educational Sovereignty examines how dependence on vendor-controlled systems can transfer educational authority to their owners (Moravec, 2026a). A school may gain sophisticated tools while losing influence over which knowledge those tools recognize and how they represent a learner’s ability. Educational sovereignty requires schools and their communities to have a meaningful say in those decisions, backed by the expertise and resources to develop or obtain systems that serve their educational purposes.
Teachers should be able to help develop and adapt these tools. Their knowledge of learners should influence how a product responds to students and what it treats as evidence of learning. Asking teachers for feedback means little if their recommendations cannot influence the design. They also need discretion over whether and how to use a tool in their teaching.
Exercising that authority requires technical and legal expertise, money, and negotiating power. Many districts cannot assemble enough of these resources alone. Katherine’s suggestion that schools could work together regionally connects with the “missing middle” in Unprepared. Regional education cooperatives, library networks, or public-interest partnerships could employ specialists to evaluate systems, negotiate contracts, and develop shared services.
For example, districts could negotiate the right to export student records in usable formats and require advance notice of consequential changes to a product. Shared technical support could make changing providers affordable. These are practical conditions for exercising public authority. A committee that gathers teachers’ views but has neither the budget nor the authority to act on them will accomplish little.
When human attention becomes a premium service
In Unprepared, I examine the possibility of an analog retreat: a deliberate reduction in technological mediation in favor of more direct human contact (Moravec, 2026c). Who will be able to afford it? Well-resourced schools could preserve sustained contact with teachers while financially constrained schools replace more of that contact with automated assistance. Both might advertise personalization. The experience of being a student could differ profoundly.
Baumol’s (1967) analysis of unbalanced productivity growth helps explain the pressure. When productivity and wages rise in some sectors, labor-intensive services face pressure to raise pay even where comparable productivity gains are difficult. Their costs rise relative to those of more easily automated activities. In education, sustained attention to a learner takes time. Increasing the number of students a teacher serves may lower the cost per student while changing the quality of the relationship.
The consequences depend partly on how schools spend any savings. If AI reduces administrative work, a school could give teachers more time with students. Under financial pressure, it might reduce staffing instead. Those are budget decisions with very different consequences for students.
I called this the “Netflixization” of education. I do not want a relationship with a teacher to become the educational version of a premium subscription.
Automated assistance also depends on physical infrastructure. Katherine raised AI’s energy demands, a concern examined by the International Energy Agency (2025). She asked us to consider futures in which climate pressures and unreliable infrastructure make low-bandwidth or offline learning necessary.
Choosing fewer screens and losing dependable electricity are very different circumstances. A screen-time measure could obscure the distinction. We need to ask who has a choice, what replaces the technology, and whether students still have people they can rely on.
The intellectual work of teaching
A school’s ability to change depends heavily on the judgment it permits its teachers to exercise. Katherine described change as happening person by person. A policy can be adopted at a meeting; the people expected to carry it out still have to understand it and work out what it means for their students.
Teachers are improvisational performers. They notice the student who has misunderstood the question, and the student who understands it well enough to ask a better one. They decide whether an unexpected contribution deserves the rest of the lesson. A lesson plan cannot anticipate every such decision. Teachers have to know their subject, understand their students, and judge when a departure from the plan is worth pursuing.
As I put it in the conversation, a product becomes educational through its place in that human interaction. Schools need to judge its value by what students actually learn through using it.
Schools learn from these judgments, too, when teachers have time to discuss them and the authority to act on what they discover. A teacher may find that students understand a concept better through a community project than through the assigned exercises. Colleagues can examine why it worked and whether it is worth developing further. Insisting on uniform implementation prevents that professional inquiry.
Build a Positive Rebellion explores how teachers and learners can question school routines and build worthwhile alternatives, even where their freedom is constrained (Moravec, 2026b). Students need opportunities to ask their own questions and undertake work they consider important. Teachers need time to think together and the freedom to respond when the work takes an unexpected direction. They should be able to discuss what did not work without assuming that every unsuccessful attempt reflects poor teaching or insufficient effort.
Katherine proposed anticipating possible futures, testing approaches, and reflecting on the results. That work requires permission to discover that an assumption was wrong, along with safeguards for students and a willingness to stop an unsuccessful experiment. I argued that refusing all risk brings a danger of its own: we can preserve a failing approach across an entire system because no one is allowed to try anything else.
Companies also have to imagine a different education
Districts buy individual products through established budgets, and companies develop offerings that fit those purchasing arrangements. State accountability requirements further restrict what a district can change. The audience’s questions about funding and piecemeal purchasing made clear how difficult it can be to work beyond those constraints.
A company can respond rationally to all those incentives and still reinforce a form of schooling that serves learners poorly. Calling on schools to transform themselves therefore places an obligation on providers to reconsider their own practices, too.
Katherine urged companies to understand the educational purposes of the places they serve and develop solutions with educators. Her discussion of modular public education raised questions about how those purposes could be pursued across several providers. If students learn across several settings, they will need adults who help them connect those experiences and recognize what they have learned. Dividing provision among more organizations also makes responsibility easier to evade. Who notices when a student’s education consists of disconnected activities?
I suggested that industry associations could help districts examine their futures without every discussion becoming a pitch for a particular product. Companies could reduce teachers’ administrative work or help schools organize community mentorships. Providers would need to cooperate so that teachers and students can work across their systems without repeatedly entering the same information or losing access to previous work.
The same questions apply to companies building on someone else’s AI infrastructure. How much control do they have over the systems on which their promises to schools depend? A supplier exposed to another company’s pricing, model updates, or terms of service has dependencies of its own. Those dependencies give schools and suppliers a reason to cooperate on alternatives. They also make it essential for suppliers to distinguish what they control from what they can only promise to manage.
Companies that take educational sovereignty seriously should expect schools to adapt products, combine them with other resources, and sometimes discontinue them. Contracts and product design should accommodate those decisions.
Keeping education public as its forms change
Katherine closed by asking how education’s role as a public good could be preserved as its structures change. That question becomes especially important when responsibility is divided among more providers.
Young people could benefit from learning in a wider range of settings. But families differ enormously in their ability to find opportunities, arrange transport, evaluate providers, and pay for services. Public education must ensure that children receive a good education regardless of those differences. Offering more options does not by itself ensure that every child can benefit from them.
Automated assistance cannot assume that public responsibility. A child still needs adults who know them and a community in which their contributions matter. Decisions about staffing, infrastructure, and the ownership of educational systems will affect whether those relationships are readily available or increasingly reserved for families who can pay.
I am grateful to Michelle, Katherine, and the participants for pursuing these questions beyond the usual discussion of what schools might buy next. Their questions made clear how much depends on decisions about funding, purchasing, and professional autonomy that are often treated as merely administrative.
We cannot promise children that we know the world they will enter. We can give them more experience of influencing the world they inhabit. We cannot prepare people to change the world by making sure they never change the school.
References
Baumol, W. J. (1967). Macroeconomics of unbalanced growth: The anatomy of urban crisis. The American Economic Review, 57(3), 415–426. https://www.jstor.org/stable/1812111
Crabtree, M., Swanson, J., King, K., Prince, K., & Righteous-Rogers, J.-A. (2025). Charting a new course for education. KnowledgeWorks. https://knowledgeworks.org/resources/forecast-7-charting-new-course-education/
International Energy Agency. (2025). Energy and AI. https://www.iea.org/reports/energy-and-ai
Moravec, J. W. (2026a). Beyond the digital enclosure: AI, coloniality, and the pursuit of educational sovereignty. Learning Futures and Emerging Technologies. Advance online publication. https://doi.org/10.1108/LFET-01-2026-0004
Moravec, J. W. (2026b). Build a positive rebellion: Create new education futures. Education Futures. https://doi.org/10.5281/zenodo.18744715
Moravec, J. W. (2026c). Unprepared: Education beyond 2030. Education Futures. https://educationfutures.com/unprepared-education-beyond-2030/
