Rethinking Education in the AI Era: Five Directions

Theme: Agent Engineering

Published:

Last Updated:

Rethinking education in the AI era

When answers become easier to find, what kind of people do we hope our children will become?

For a long time, education has revolved around acquiring knowledge, answering correctly, and passing selection exams. Students follow a relatively uniform curriculum and advance through successive examinations. This system has given many children access to basic education. It has also deeply shaped our idea of excellence: learning quickly, answering accurately, and earning high grades.

AI gives us a chance to ask again: what should a child gain from more than a decade of education?

Knowledge still matters, and fundamentals still require practice. But when students can access explanations, feedback, and creative assistance whenever they need them, many familiar arrangements deserve another look.

Here are five directions I hope education might take. They may never fully materialize, but I believe they are worth looking forward to.



1. More effective learning could make personalized pathways possible

Traditional classrooms face a persistent tension: one teacher must guide children with different foundations, interests, and learning speeds toward shared goals within limited time.

Grouping students can reduce broad differences, but within the same lesson some already understand while others are stuck on prerequisite knowledge. Teachers must consider the whole class, making it difficult to give every child the right challenge.

AI opens up new possibilities. Students can ask about the same concept repeatedly, request a different explanation, and receive feedback tailored to their mistakes. They can revisit what they have not understood and reduce repetition where they are already proficient, leaving time for deeper questions.

A child who reads well but struggles with mathematics need not be reduced to a total score. They could receive patient support in mathematics while continuing to explore through reading. A student progressing slowly need not see themselves as someone who cannot learn simply because they missed a shared milestone.

But learning efficiency requires a clear view of the goal. Finishing exercises faster does not necessarily mean understanding more deeply. Completing a task with AI does not prove that a student can perform independently. What can be accelerated varies by age and subject. We need to observe whether children can explain, transfer, and independently apply what they learn, alongside how quickly they finish.

Better personalized education should give children appropriate support while leaving room for possibilities beyond existing assessments.

Some parts of knowledge acquisition can become faster. Developing judgment, understanding oneself, and learning to live with others still require experience and time.

That raises the next question: if we save some time, what should we give it to?



2. Redesign the school day to make room for life skills and practice

Suppose a student can master the same academic material in less time. What happens to the rest of their day?

My fear is that they finish this year's curriculum only to start next year's early. Content moves forward, competition follows, and greater efficiency becomes a denser workload.

We could make another choice: give children time for sport, reading, creative work, real projects, learning to care for themselves, and accomplishing something with other people.

Many abilities we want children to develop have long been squeezed into extracurricular time. Managing time, handling disagreement, expressing ideas, honoring commitments, and facing failure profoundly affect how a person lives.

The school day deserves redesigning. Academic study, sport, art, practical work, and community participation should each receive time that reflects their value. Children need different activities through which to discover themselves and connect knowledge with life.

Field learning offers one opening. Students might investigate difficulties older residents face with digital services. They need to prepare, find people willing to talk, understand their circumstances, and then propose improvements.

They may discover that the problem they imagined does not exist, while a real difficulty escaped their attention. A technically workable solution may be something people have no desire to use. Those setbacks are precisely where learning happens.

AI can help organize material, analyze notes, and explore possible solutions. Listening, observing, and responding in person remain the students' responsibility. Real participation gives them a chance to develop personal capacities beyond academic knowledge.



3. Put building and motivation at the center so students can initiate learning

Many students know one sequence well: follow instructions, learn the material, complete assignments, and receive an evaluation.

That path provides necessary structure. But when it becomes the only path, children may become increasingly good at completing tasks while rarely asking: what do I want to make? What do I want to understand?

AI lowers the barrier between an idea and an initial result. Students can try making a small tool, producing a video, researching a local problem, or organizing an event. Ideas previously set aside for lack of skills now have a starting point.

When students want to make something work, knowledge gains a concrete purpose. They revise writing to explain an argument, learn programming and data to make a tool useful, or discover that understanding survey results requires statistics.

A goal drives learning, and learning supports action. Through that cycle, students can begin to experience how their effort creates real change.

Building should have a broad meaning. Studying an insect, rehearsing a play, or improving a classroom library's lending system all deserve serious attention. Education should accommodate different interests without steering every child toward entrepreneurship, coding, or a single image of success.

Building also requires foundational learning. A polished result generated quickly by AI can hide gaps in understanding. Teachers need to ask children to explain key choices, change a condition, and face a new problem to establish whether their abilities are developing.

Motivation does not mean constant enjoyment. Meaningful work often includes boredom, frustration, and long stretches without results. Education should help children break difficulties into smaller parts, find appropriate support, and gradually learn persistence.

A worthwhile goal is to nurture someone who can initiate learning: a person with questions they want to understand and the willingness to act on them.

That autonomy grows gradually. Younger children still need clear arrangements and adult companionship. Choice and responsibility can expand as their abilities grow. Handing everything to an unprepared child does not automatically produce initiative.



4. Let educators devote more attention to human growth

AI can assist with explanations, exercises, and marking. Teachers' work may change, but some of education's hardest problems remain.

Is a child silent in discussions because they do not understand, or because they fear being mocked? Does a falling grade indicate a knowledge gap or something happening in their life? When a student asks for help, should a teacher offer a hint or wait patiently?

Such judgments require knowing a particular person over time.

Educators will increasingly need to design meaningful tasks, observe students, organize collaboration, and offer support when failure, conflict, and confusion arise. Teaching knowledge remains important. Work centered on people deserves more time and respect too.

This could give human qualities and systems thinking a larger place in education.

Systems thinking helps students see connections. Improving a school cafeteria queue involves space and processes, different year groups, staff workload, and student habits. A seemingly efficient solution may transfer costs to someone else. Teachers can help children ask: who is affected? What other factors matter? What consequences might follow?

Grades provide some information. Understanding complex problems, collaborating with others, and taking responsibility for the consequences of one's actions should also inform how we assess growth.

If the time saved by tools merely allows one adult to manage more students, educational relationships may become thinner. Technological efficiency needs to become attention and support that children can actually feel.



5. Make schools hubs in a wider learning network

Students may gain access to a wider range of learning sources: courses across schools, remote mentors, specialist communities, neighborhoods, museums, laboratories, and real projects.

A child interested in astronomy might find enthusiasts and researchers willing to guide them. A student who loves theater could join a production across schools. A group tackling a local issue could work with community organizers. Beyond school, many people and experiences are worth connecting with.

Decentralization here means that knowledge and opportunities come from more places. It also demands stronger coordination from schools.

Scattered resources do not automatically make a good education. Students need help selecting opportunities, judging their suitability, connecting experiences, and receiving sustained support when difficulties arise. For minors especially, open learning networks also require reliable safeguards.

Schools could gradually become hubs: providing a stable shared life while helping students reach a wider world. Children would have places to explore and somewhere to return for conversation, reflection, and support.

This direction raises questions of fairness. Families with more resources can more easily find mentors, projects, and practical opportunities. Access for other children cannot depend solely on their families' time, knowledge, and connections.

As learning becomes more networked, schools need to take greater responsibility for connecting opportunities. Children who would otherwise struggle to reach specialist resources should have a chance to be seen, guided, and involved in worthwhile work.

The value of a wider educational network should be measured by the opportunities it makes available to more children.



A final thought

Ultimately, this path concerns how a person grows: understanding the world, discovering themselves, initiating learning, participating in reality, living alongside others, and gradually making choices about their own life.

Technology does not guarantee that outcome. The same AI can increase exercise density, bring competition forward, expand surveillance, or attach more detailed labels to children. What tools can do and what education chooses to do remain separate questions.

We need to decide again what deserves assessment, what deserves time, and what every child deserves to retain.






Related reading

A personal skill map for the AI era