The end of education as we know of

We are approaching the end of education as we know of.

There are two types of training that mankind has offered for generations.

  • Training for knowledge
  • Training for critical thinking

The first type of training has become and will continue to become useless as it can be replaced by ChatGPT like LLMs, or AI.

The 2nd type of training is still human teachers’ responsibility, but the delivery is challenged by students over-relying on LLMs.

For the knowledge delivery, I have raised two thoughts.

Training for knowledge

1.The authority of knowledge: Guru → Library (books) → Google → AI

For Gen Alpha and Beta, the LLMs are like Google for Millennials. Back in the days when knowledge was delivered by teachers who relied on books written by Guru, students had to go to libraries to find the original references.

Since 1990s and early 2000s, the Millennials have adopted different practices. Instead of going through books in libraries, the Millennial generations have relied on Google to find information. For me, to find an academic reference, the first source of reliable search is Google Scholar. Yet, baby boomer professors had resisted to adapt to the reality, and some of them asked us to find the authoritative information from books in libraries.

For my generation, such a professor is not just an asshole. Boomers for us are no more than just monuments and fossils.

Now, if we impose certain rules on using when and where to use AI tools in class, Gen Alphas and Betas will see us the way we saw boomer professors.

Given the fact that these keys have to be conversant with AI tools on the very first day of their job, refraining them of the vastly advanced tool for study won’t be an effective teaching anymore. You can’t train your students in 2020s the way that you were taught in 2000s.

2. Teachers are losing their authority over AI

Back in the days when I was a student, when textbooks are difficult to understand, the first source of alternative explanation was Google. It’s not completely reliable, but many external sources helped us to understand the materials better.

In 2020s, students go for LLMs to find backup materials when they can’t understand the lecture materials.

What’s more interesting is that many students no longer spend time for studyting. They just copy and paste answers from LLMs, hoping that the copied answer is immune from hallucinations.

Back then when my generation copied Wikipedia’s proof for a math proposition, we knew that we had to adjust the proof for the purpose of the assignment and the entire curriculum, which often is not 100% synchronized with what Wikipedia (or even textbook) says.

But, in 2026, LLMs provide us answers that look so closely adapted to class materials, and many of them actually are good enough to pass the evaluators’ keen observation.

Without LLMs, we still had to digest the computer’s output, although the degree might be less intense than the generations who had to go to libraries.

With LLMs, the new generations do not have to think too much about computer’s output. Because most of them are right.

What’s more interesting is, when there is a discrepancy between teachers and AI, students begin to rely more on AI over teachers.

Why? Teachers refine the scope of teaching for that class’s purpose. AI tools do not impose such a limit.

It’s like, in Linear Algebra, professors teach eigen decomposition through matrix transformation, but students find solutions that are based on Topology. Considering the fact that Linear Algebra is mostly the first year STEM course and Topology is 3rd to 4th year courses at Math department, it’s difficult for professors to use Topology in the 1st year class without sacrificing some slower students. But, students equipped with AI tools are no longer bound to that limitation, although many of them do not even have taken pre-requisites for Topology. Because of that, the situation gets worse. Sometimes professors have to stop progress in class and teach Mathmatical Analysis, which is the first course in the 3rd year Math department and which is the foundation of all higher level math courses like Topology.

With Wikipedia, I remember we already had similar problems, but the degree wasn’t intense.

With LLMs, the degree has become much more serious.

Training for critical thinking

As much as students offload their cognitive ability in training for knowledge, teachers will likely also be in the similar situation, because deliverying knowledge isn’t something that requires peculiar skills.

Increasingly, we will see that what’s left for us is the training for critical thinking, but I see even more resistance from students. They already see that college education no longer pays off. MOOC based education is just as good or sometimes better, when it comes to training for knowledge.

For teachers to enforce and emphasize training for critical thinking in classes, there has to be certain requirements by the social system.

  1. Employers require critical thinking from job candidates
  2. Teachers are ready to teach ‘how to think’

Employers require critical thinking from job candidates

Let’s think about Korean securities companies for the requirement of ‘critical thinking’.

Quick disclaimer: I have never worked in Korean companies in my life, so this story is based on 2nd-hand experiences, but I believe it is close enough to be the reality in Seoul.

Back in my college days, many friends in the business department often complained about lack of ‘hands-on knowledge’ or ‘real knowledge’. Sometimes they also said ‘money-making knowledge’ and ‘corporate knowledge’.

I hated them from the bottom of my heart, and I still do. Why?

For a corporate finance class, for example, we learn M-M theory to understand how the capital choices are related to financial valuation, which also coincides with the market valuation in theory, but in practice, the market’s daily reactions often have more information than the accounting books describing past information.

If one is trained in the way I hoped, the educational output will have people creating following interpretation

and, readers digesting the information in similar manner.

But, if the school’s training is about mechanical knowledge to earn more scores, like Korean education, hardly anyone cares about the aforementioned article’s insight. It doesn’t help them to earn more score in the exam, and it doesn’t help them to get a job in Korean financial market. Then why bother?

Those business major students graduate with deep distrust to school’s education, and they just think about school as for network building.

The consequence?

Koreans invest in stock markets based on guts and rumors, not based on analysis and criitical thinking.

The failue in college education resulted in financial market’s instability, the cost of which is country’s risk premium.

In other words, those business major students without critical thinking got jobs in financial sectors because Korean brokerage firms do not have rigorous vetting in the level of critical training, which results in nation’s major financial cost. Haven’t checked in detail, but I am sure that it is more than just 1% in equity premium that the Korean market has to compensate to global investors. 1% in equity premium multiplied by the sheer amount of FDI is already multi-billion dollars.

Teachers are ready to teach ‘how to think’

The true nature of the Korean financial market problem is the lack of proper training for critical thinking, not just at a college’s business department , but at the society as a whole.

Korea is known for the highest per capita investment in education for every kid, but the output has been criticized due to under-appreciated quality. For example, despite 12 years of English teaching, hardly any Korean properly speak English. If you speak good English in Korea, people often assume that you’ve lived in English speaking country for your childhood, instead of doing good at school. Why? The school’s English education is focused on getting highter mark, not training students conversant in English.

When it is not the true ability but the exam’s score that determines each individual’s job candidacy, cheatings become rampant. Korean Hakwons, private tutoring mostly focused on getting higher mark in exams, have become the central component of Korean private sector education, largely because the society as a whole relies on exam marks, instead of multi-dimensional evaluation of one candidate’s true ability.

Teachers are trained by that system. Every corporate seniors are trained by the same system. Now, who’s going to teach younger generation how to critically think?

When they find jobs, even in the age of AI, they are still evaluated by the exam scores. Would they care about critical thinking?

After they get the job, no senior in workplaces is ready to give me proper corporate training with critical thinking.

The degree might be more intense in exam-driven society like Korea, but I think the overall situation globally isn’t that different, when it comes to lack of teachers for critical thinking.

Many boomers and millennials have found their jobs with mechanical knowledge, which can be done better by LLMs.

Out of 100 teachers, less than 10 of them are good at teaching critical thinking. Others? They have the teaching job by delivering knowledge, mechanically.

The end of education as we know of

Soon, schools and universities will lose its market recognition. Students will likely go to MOOC and online alternatives to find necessary survival skills. Places that foster critical thinking, if corporate HR choices are matched, the value of their education will be more emphasized than now. If companies hire people just like Korean companies, then the country will lose critical thinking education entirely.

I have been pessimistic about Korea, and having seen what educators say at Bett Asia, I become more convinced. And, unfortunately, the situation won’t be that different in other countries as well.

Except very few educational institutions that are capable of training students for critical thinking, other places will continue to be replaced by MOOC, which one day might be fully automated by AI.

At SIAI, we already make such a system for our own education. Others will soon follow, Because teaching is, ironically, the most stressful and time-consuming job for professors.