As AI Gets Smarter Than Students, What Should Schools Teach?

Cheating with smart glasses and generative AI is the immediate problem. The deeper one is that grading finished answers no longer shows what a student understands.

The 20-second read

1
Tiny devices and AI are shaking trust in exams, but tighter proctoring alone cannot fix the problems in education.
2
The OECD points out that doing well on assignments with AI is not the same as actually learning.
3
Going forward, the goal of schools should expand from producing correct answers to building independent thinking, collaboration, judgment, responsibility and a sense of direction in life.

In November 2026, cheating with tiny cameras and generative AI came to light in a national licensing exam in South Korea. Finding smart glasses, earpieces and communication devices is the immediate task. Proctoring and rules should be tightened to keep exams fair. But ending the discussion there misses a bigger problem.

A bigger problem than cameras in the exam room

If more students submit work they cannot explain, that is a failure of assessment as much as a failure of security. Schools have long graded finished answers. Now they must separately check how far an answer reflects the student's own understanding. Once a machine could produce correct answers, judging human ability by correct answers alone began to crack.

What calculators and spreadsheets changed, and what AI will change

Calculators in the 1970s sharply reduced hand calculation. VisiCalc, introduced in 1979, cut the repetitive arithmetic in ledgers. Yet calculators did not eliminate math, and spreadsheets did not replace financial judgment. As tools improved, the ability to decide what to calculate, which assumptions to set and whether results match reality became more important.

AI differs in scope. It does not only calculate. It drafts questions, structures essays, writes code and even offers counterarguments. In future, machines may handle not just the output of an assignment but the step of defining the problem. Should education then race to find what technology cannot do? A better starting point is to ask which abilities humans must own for themselves.

A well-made assignment is not the same as learning

The OECD's 2026 digital education outlook concluded that better task performance with generative AI does not automatically translate into learning gains. A student may turn in an excellent report with AI's help yet be unable to explain the same principle, or apply it to another problem, without that help.

A 2026 study of college writing found a similar pattern. Free collaboration with AI raised the quality of submissions. But the group that had AI use limited and was asked to reflect tended to do better on later independent writing tasks. This was a quasi-experiment comparing a small number of classes, so it cannot be assumed to hold for every student.

The problem is the outsourcing of thought. People have always handed memory to notepads and arithmetic to machines. Useful outsourcing leaves time for more important thinking. Harmful outsourcing skips the thinking process that was supposed to be learned. A fluent AI explanation does not become the student's own understanding.

Why education needs some deliberate difficulty

Learning does not end when information appears on screen. Knowledge becomes one's own through retrieving memory unaided, correcting wrong hypotheses and applying ideas in new contexts. That is the point of cognitive psychology's "desirable difficulties" and "productive failure."

That does not mean making students struggle for its own sake. Difficulty must match the learner's level, and failure must be followed by explanation and feedback. An AI tutor may serve education better if it asks for a student's first attempt, offers hints when needed and finally has the student explain without help, rather than showing the answer at once.

Handwriting and paper books are one option. But claims that analog methods are neurologically superior to digital ones for every student and task should be avoided. What matters is less the medium than whether the student actively processes the content and retrieves it from memory.

Redesigning exams, starting at US universities

Researchers at NYU's Stern School of Business tested oral exams in which a voice AI asks follow-up questions about a student's project. In the initial study, 36 exams were run for roughly $15 in total model costs. Seventy percent of students said the format assessed real understanding, but 83% found it more stressful than a written exam. The results show both the promise and the limits of oral assessment.

Oral exams are no cure-all, whether a teacher or an AI asks the questions. Speaking anxiety, language ability, accessibility for students with disabilities, grading bias and privacy all need attention. So separating assessments is more realistic than abolishing exams. Check basic concepts without AI, allow AI for real problem solving, and finally have students explain why they made their choices and what the limits are.

What might school look like in 2031 to 2036?

AI performance, cost, regulation and public acceptance over the next five to 10 years are uncertain. No one can promise a perfect personal tutor for every student. Still, a scenario in which AI tutors that gauge each student's understanding and adjust questions and feedback are used more widely than today deserves serious consideration.

Should the classroom then remain a place that delivers the same content at the same pace? If AI takes over part of explanation and repetitive practice, teachers could spend more time helping students understand why they learn, spotting false confidence, resolving conflicts among friends and encouraging them to try again after failure. But if AI is adopted only to cut costs by reducing human relationships in education, the opposite could result.

Imagine a day in that future. In the morning, students read books and prove mathematical principles without AI. In the afternoon, they use AI to investigate local transportation, energy and care problems. They verify data sources, meet real people and balance competing interests. At the end, they explain their decisions to teachers and peers. This is not a settled future. It is a design scenario that shows a direction for education.

The easier knowledge gets, the more basics matter

If AI can explain everything, can memorization and foundational learning be dropped? The risk runs the other way. People who lack basic vocabulary, number sense, historical context and scientific principles find it hard to judge why an AI answer is wrong. They may not even think of what to ask.

So future education should not abandon foundational knowledge and teach only creativity. It must connect internalized basics to higher-order use. Reading, writing, math, science, history and civic education remain important. What should change is the habit of treating how much knowledge a student can reproduce as the only measure of success.

What remains for humans in the AI era?

If AI takes over parts of writing, analysis, design and counseling, what happens to human jobs? It is more prudent to see tasks within jobs being reorganized than to assume particular occupations vanish entirely. Humans will likely keep setting goals, mediating conflicting interests and taking responsibility for outcomes, though AI will affect those roles too.

A more basic question remains. Should human worth be judged by productivity alone? Caring for a child, comforting a friend, taking part in a community, sensing beauty and choosing the direction of one's own life are hard to measure in economic output. If education focuses only on producing people competitive in the labor market, then the more AI raises productivity, the more anxious people may become about their reason for being.

So schools of the future need to teach purpose along with ability. That means asking not only what a student can do but what the ability is for, whom it affects and how to take responsibility when things go wrong. Ethics, philosophy, art, physical education and community experience become central to forming a person, not the periphery of admissions.

What to change first

First, change assessment. Separate AI-free exams from AI-enabled assignments, and in every subject check independent performance and the ability to explain, not just the product. Second, clarify the purpose of homework. If it is meant to check understanding, it must be designed so AI ghostwriting does not work. If the aim is real problem solving, the use of AI and the verification process should be assessed together.

Third, reallocate teachers' time. AI can assist with some grading and repeated explanation, but students' emotions, motivation, relationships and safety should not be left wholly to machines. Fourth, narrow gaps outside school. Public education should provide fair access and guidance so that paid AI tools and home support do not decide grades.

Fifth, college admissions and hiring must change too. If schools teach reasoning and collaboration but universities and employers still reward only speed to the right answer, education slides back into a contest of problem solving. Institutional change has to move together with change in the classroom.

If AI gets smarter, is there any reason to memorize knowledge?
  • Basic knowledge is the foundation for checking AI's answers and asking good questions. Neither expanding rote memorization across the board nor scrapping basic knowledge altogether is the answer.
Are exams that ban AI outdated?
  • They are needed when measuring independent understanding. But to assess real problem-solving skills, there should also be separate assessments that allow AI.
Can AI oral exams replace human teachers?
  • They can help with repetitive follow-up questions, but fairness, emotional safety, appeals and final responsibility still need human oversight.
Will there be fewer teachers in the future?
  • Administrative and explanatory tasks may change. But demand for teachers depends on policy and budgets, class size and social choices, so it is hard to predict from technology alone.
What can parents change starting today?
  • A good starting point is to have your child state their own thoughts before opening AI, and, after getting an answer from AI, help them explain what was wrong and why they chose to accept it.

What to watch

Achievement gapThe gap in results between independent performance assessments taken without AI and assignments that use AI.
Oral assessment and AI tutorsThe reliability of oral assessments, accessibility for students with disabilities, and student stress levels. Also, changes in long-term memory, transfer ability and self-directed learning after AI tutors are introduced.
Access and admissionsDifferences in AI access between schools and learning gaps by family background. Also, how much weight college admissions and hiring give to explaining one's process and to real problem-solving.
When AI can produce the answer, assessment has to shift toward checking how well the student understands it.

Insight Times Editorial Desk

More by topic

Starting with this story's topics, then what readers are following now

Investor psychology, decision-making and books on investing

GenAI

This story's topic15 storiesSee all →

AI that produces something new — text, images, audio, code — rather than class…

Box Shows AI Its Manuals, and the AI Starts Redesigning the Workflow

Box Architect, built with NVIDIA, moves AI from reading documents to designing how work gets done. It has not yet reached beta, so the real-world payoff remains to be seen.

Tech

China's AI Labs Went Open Source Because They Had to, Not Because They Wanted To

A GPU shortfall pushed Chinese AI labs toward efficiency and open weights, and that constraint is now reshaping the global competition over who controls AI's basic infrastructure.

Tech

2027 Investing Needs Alignment, Not Prediction

A new investing book argues the hard part was never picking the next winner. It's staying aligned with where value and bottlenecks are moving while keeping old human weaknesses in check.

Investor Mind

AI Competition Shifts From Chatbot Share to the 'Action Layer'

ChatGPT isn't collapsing. But as Gemini, Claude and Copilot grow through their own distribution channels, the real contest is moving from standalone chatbots to an action layer wired into search, operating systems, workplace tools, ads and payments.

Tech

The Number That Matters More Than How Smart AI Gets

This week's AI models got stronger and cheaper at the same time. The chart long-term investors should watch isn't the weekly benchmark leader, it's the falling cost of finishing a task successfully.

Tech

Can AI-Written Text Now Be Caught? ChatGPT's Invisible Watermark and Its Limits

OpenAI has begun embedding an invisible watermark in AI-written text, following Claude. The signal can suggest AI involvement, but it can't say who is accountable for the writing.

Tech

AI's Next KPI Isn't Token Price. It's Cost Per Successful Task

Apple is cutting the cost of where inference happens, Anthropic is cutting the cost of running its models, and Snorkel AI shows money flowing into the data and evaluation layer that reduces failed attempts. All three point to one shift: AI competition is moving from bigger models to cheaper successful outcomes.

Tech

A building that houses servers and networking gear and supplies them with powe…

Why Nebius Is Keeping Some 2027 AI Capacity Off the Shelf

Nebius mixes long-term contracts, short-term premiums and auctions. Holding back capacity pays only if the servers stay busy, and it is not yet clear the pricing will last.

Tech

AI ran out of memory before it ran out of GPUs

The HBM bottleneck is spreading into DRAM prices, server costs and the price of the next laptop or phone.

Tech

Is Space Cooler for AI? Why Musk Wants Data Centers in Orbit

The next limit on AI may be power and cooling, not GPUs. Elon Musk is looking to Earth orbit for an answer, but the test is cost per hour of compute.

Tech

White House Swaps AI for "SI." The Real Investment Signal Is Infrastructure, Not the Name

An executive order renames artificial intelligence "Super Intelligence" and six frontier labs sign a voluntary safety accord. What investors should watch is the data center, power, nuclear and chip policy now gathered under one national-competitiveness frame.

Tech

Musk Sees the Next Phase of AI Infrastructure: Who Can Switch GPUs On, Fast

If neoclouds struggle under financial pressure, AI demand may not shrink. It may simply favor the few operators with the capital and execution to serve it, a shift that could matter for SpaceX and Tesla.

Tech

Weekly Check: The AI Bull Market Isn't Over, It's Just Gotten Harder

A 5% 10-year yield, 470 gigawatts stacked up in Texas, Chinese DRAM, and whether data centers actually get switched on are starting to decide who wins the AI trade through 2027.

Tech

Is Memory Peaking? In the AI Era, Watch the Bottleneck, Not the Demand

Demand still outruns supply. But memory stocks can peak before demand does. The real question is how long HBM4, server DRAM, CXL and Big Tech capex keep the cash flowing.

Tech

NVDA

101 storiesSee all →

Who Hits $1 Trillion Next: Reading AMD and ASML by Different Rules

One company is chasing Nvidia's GPU empire. The other sells the machine that draws the chips for every advanced chipmaker, including Nvidia. Both get floated as 2028 trillion-dollar candidates, but the numbers investors need to watch are entirely different.

Investing

Tech Selloff: Is the Bull Market Over? Wall Street's Four-Year Chart Says Not Yet

The uptrend has survived five double-digit pullbacks in four years. This time, AI companies' debt and cash flow are the new variables.

Investing

Marvell's $115 Billion Lumentum Deal: Is AI's Next Bottleneck Light, Not Chips?

Marvell's purchase of Lumentum points to a contest beyond GPU speed, in data transfer and power efficiency. Whether the price is right is not yet confirmed.

Tech

Nvidia and Palantir Put AI Inside the Supply Chain - The Real Shift Is That It's Learning Judgment

A single Vera Rubin rack holds 1.3 million parts. The real breakthrough in Nvidia and Palantir's new supply chain AI isn't automating that complexity, it's capturing the judgment calls that used to live only in planners' heads.

Tech

If You Only Watch Tesla, You May Be Missing Half the AI Story

Nvidia builds the AI brain. Who supplies the electricity to run it? Six power companies Tesla investors should know, and why Tesla's storage business is just one link in the chain.

Investing

Nvidia Moves From GPUs to CPUs as AI Agents Reshape a $200 Billion Chip Market

A customer's early test found Nvidia's Vera CPU handled about twice the AI agent workload of its existing chips. The bigger story is where the money in semiconductors is heading as AI shifts from answering questions to doing work.

Tech

Nvidia's $500 Billion Experiment: Wall Street Asks Who Backs the GPU, Not Just the Chip

The real issue in Nvidia's financing platform is not chip sales. It is how far Wall Street will trust GPU resale prices and rental income three to four years out.

Tech

GPU

73 storiesSee all →

A chip built to draw graphics, whose ability to run thousands of identical cal…

The Next AI Chip Race Is About How Chips Connect

Applied Materials will lift its stake in Dutch equipment maker BESI from about 9% to 20%. The bet is less about control than about a new way to link chips.

Tech

US AI Data Centers Are Headed for 90GW. Gigawatts Now Matter More Than Dollars

Goldman Sachs expects US data center capacity to grow from 64GW at the end of 2026 to 90GW at the end of 2027. The AI race is moving from buying GPUs to building power plants and transmission lines alongside them.

Tech

The Cheaper Tokens Get, the Bigger Your AI Bill Gets

As AI shifts from training to inference, the token is becoming an industrial commodity. Falling prices are not cutting demand, because agents and automation consume tokens at a scale no human chat session ever did.

Investing

Dark navy cover. Flow diagram: Nvidia to Nscale (credit guarantee, in talks) to Anthropic (~$45B compute lease signed in August plus a separate ~$45B, ~460MW deal in talks). Title: Why Would Nvidia Back a Customer's Loan?

Why Would Nvidia Guarantee Its Customers' Data Center Loans?

Nvidia is reportedly in talks to back financing for Nscale's West Virginia data center. Nothing is signed, and the terms are not public.

Tech

Two Companies Drive 34% of S&P 500 Earnings Growth as AI Pays Sellers First

Nvidia and Micron generate a third of expected earnings growth in US large caps. Apollo says the financial effect of AI is not yet visible in the S&P 493.

Investing

Amazon Goes After Nvidia's Strongest Moat, and It Is Software, Not GPUs

Amazon's open-source TorchNeuron aims to let existing AI code run on its Trainium chips. The contest comes down to the total cost of AI compute and the return on Amazon's $220 billion in annual capital spending.

Tech

AI's Next Bottleneck Is Power, Not GPUs

Jensen Huang and Elon Musk sent the same warning in Washington. The AI race is widening from building better models to supplying power and chips at industrial scale.

Tech