Evolving Role of Mentorship

Are you an educator?

What are your thoughts on Large Language Models, and AI technology in general, with respect to education? 

Do you think that soon (or very soon) you will be replaced?

Have you tried NotebookLM AND guided learning from Google or OpenAI?

If not, try them and answer again.

What do you think now?

From books to AI

New technologies have evolved human knowledge transfer before, slowly removing the need for direct human-to-human connections. From the beginning. learning required mentorship, in a form of a direct, personal transmission of skill and wisdom from an elder to an apprentice. This fundamental model, though transformed by successive waves of innovation, has never lost its core importance. However, it can be the case that in an AI-driven educational landscape, the role of the mentor, the teacher, can be diminished. Or, so they say…

Well, what if… here me out! What if, it will not be rendered obsolete; rather, it will become more critical and complex than ever before?

Johannes Gutenberg invented the movable-type printing press (around year 1440).

The first major shift in the education came with formalized institutions like church schools, which introduced the "teacher" as a figure responsible for instructing groups. The invention of the printing press democratized knowledge further, bringing books to a wider audience. Yet, mentors remained vital. Why? Because a book can provide information, but only a mentor can ignite curiosity, tailor understanding, and guide the application of knowledge.

As societies evolved, so did education. The rise of private schools for upper-caste families of all genders underscored the value of a good teacher as a social and intellectual catalyst, one who could give their students a distinct advantage. With the advent of universal public education, the teacher's role expanded dramatically again, now tasked with navigating the diverse needs and backgrounds of almost entire populace.

Not the add but unsolicited advice, read the book or watch lectures. Scratch that, do both and visit Dr.Sandel seminar whenever you can!

In the modern era, we have come to understand that nearly anyone can learn (that was not obvious at first!), provided their individual needs are met. This places the mentor at the center of the learning process again. They were required to understand and apply proper tools to accommodate any learning style. The once-underappreciated roles of kindergarten and elementary school teachers are now rightfully seen as foundational, shaping not just knowledge but a child's entire disposition towards learning. All genus engineers who are building these new tools, have got their propensity to learn and innovate, more likely from their first mentors and teachers. At the highest levels, in universities now accessible to more people than ever, the lecture of a great professor remains a coveted experience (even on YouTube! Have you watched “Justice” from Michael Sandel? Trilling, right? Can’t stop watching!) because it is more than a recitation of facts. It is a masterfully curated and condensed synthesis of knowledge, delivered with a passion that a textbook cannot replicate.

Then, the Internet threatened to change everything. With information instantly available, many questioned the need for teachers at all. If a lecture is recorded, are all professors redundant? …Probably few should stay to record more lectures, others can go, right? The reality proved more complex. The role of the professor shifted from a "sage on the stage" to a "guide on the side," helping students navigate the ocean of information and differentiate signal from noise. 

Online education tools

Online learning platforms like Coursera and edX were heralded as the next revolution, yet their impact has been nuanced. Research consistently shows that while millions enroll, the average completion rates for these Massive Open Online Courses (MOOCs) often hover in the single digits, highlighting a critical missing element: the accountability, motivation, and personalized guidance that a live mentor provides.

Yeah! It puzzles me too! But it seems to be true. 

Now, we face AI in the form of Large Language Models (LLMs). AI enthusiasts proclaim that teachers will be among the first professions to be replaced. And the best is:In addition, more individuals will be able to get education inn a more efficient way (due to AI personalization). But so far, the opposite feels true. The existence of tools that can flawlessly "cheat" a traditional assignment has not eliminated the need for educators; it has made their job infinitely more challenging and important. To truly educate a student now means guiding them through technical knowledge, and at the same time instilling professional skills like communication, leadership, and teamwork while simultaneously teaching them to use AI efficiently and ethically. It requires ensuring that genuine learning is happening beneath the surface of AI-assisted output.

Is it a surprise, that this new reality demands a significant increase in a teacher's workload? Let me show you! Gone are the days of easily graded multiple-choice questions or long essay assignments that can now be generated in seconds and, hence, are not helpful with learning. Properly designed and meaningful assessment requires more feedback, more review, and more innovative assignment design (those that can provide learning with or without the use of LLM). The idea of using AI to grade this new wave of assignments is a paradox, in my view, if an AI can grade it, an AI can do it, creating an inescapable loop of potential cheating. The path forward requires a fundamental redesign of our pedagogical approach. We have a lot of work ahead of us, not less… 

Before you live, answer these two questions:

Question 1: Who is more likely (Human educator or LLM Chatbot) to admit that they don't know the answer or 'not sure'?

Who is more likely to admit that they don't know the answer or 'not sure'

• Human Educator
• LLM ChatBot

As an educator, I know what I would do, but I was not sure about LLM ChatBot. I decided to ask one that among the most advanced at the moment of writing this. Here is the answer from Gemini 2.5 Flash:

and ChatGPT (without subscription, model that was used - ChatGPT 5)

So, here is your answer! Humans can be embarrassed by not knowing something and may try to hide it, when maintaining authority is important, while LLMs will be honest. Always.

Question 2: When you are making a mistake, who will help you by providing a correct answer with supporting explanation, and who will help you to discover the correct answer yourself?

When you are making a mistake, who will, more likely, help you to discover the correct answer?

• Human Educator
• LLM ChatBot

Guided Learning Option at Gemini

I used Gemini guided learning to solve multiple different problems from the class Transportation Systems Analysis. One you can find in tools. If you want supporting guidance while you are practicing your homework you needed to select that option.

While Gemini was very helpful, it was more similar to following practice problem in the book or professor solving problem in the class than guided learning.

Comments after first step (without mistakes)

Gemini response after the second solution step where I made a mistake

Comments after final steps, again with mistake in my calculations

This looks great! But it would not be a way how I guide the student. There are no room for critical thinking for the student, no discovery.

There is no way to stop students from using these tools to solve homework assignments, and we should not! Whenever they join the workforce, these tools will be available for them and, more likely, they will be required to be used by the employer. So, we should students how to use them, efficiently and ethically. But at the same time, we still need to teach students ho to understand the core principles of specific discipline, and that requires from us the development of new and innovative teaching approaches.

So, educators and researchers in the education, time to roll up our sleeves. We indeed have a lot of work to do!