Fixed Form, Live Attention: Learning in the AI Era

Are my cognitive skills going down because I use AI too much? This question has been doing the rounds in my peer circles for some time now… and it usually leads to a bigger one. If AI can answer most (if not all) of our questions, do we still need formal classes? Isn't AI a better disseminator of knowledge than most lecturers?

It is a fair question. In many colleges, the lecturer still walks in with notes and reads them out while students furiously copy them down. Book sales, I am told, have dropped sharply because information is available aplenty. Content now comes in every form, and moves easily from one form to another. Upload a PDF on any topic to an AI notebook tool such as NotebookLM, and within minutes you can draw out summaries, flashcards for an exam, multiple-choice questions or a mind map. And when you tire of these, you can turn the same material into an audio podcast, a debate, a presentation, or even a video… The list is endless.

So, what is the role of the educator? And, more importantly, are rote learning and knowing the fundamentals still relevant?

This is where I think we can take a leaf out of Kalaripayattu.

A tradition carried by the voice

Kalaripayattu, one of the oldest martial art forms, was once a serious method of warfare and of settling duels, especially among the chieftains who ruled the region we now know as Kerala. Under British rule, the practice was suppressed in the early nineteenth century  and it went underground, to resurface almost hundred and forty years later. Practitioners credit the vaithari (the spoken commands, called out by the master, that set the order and pace of each routine) with carrying the sequences through that dark period.

The art was, and still is, taught in four stages:

1. Meythari: body exercises that build flexibility, balance and agility
2. Kolthari: training with wooden weapons
3. Ankathari: training with metal weapons
4. Verumkai: bare-handed combat

The order matters. The student first becomes conscious of their own body. Weapon training then teaches them to focus, to be present in the present. And finally, the weapon becomes an extension of the body itself. Phillip Zarrilli, who studied and taught Kalaripayattu for decades, takes the title of his book from this ideal: the state in which the body "becomes all eyes" (Zarrilli 1998).

What the theater saw

Eugenio Barba, founder of Odin Teatret, describes the same phenomenon from the stage. In A Dictionary of Theatre Anthropology (Barba and Savarese 2006), he separates the techniques we use in daily life from the "extra-daily" techniques a performer trains, which put the body in form before it expresses anything at all. In his own company, actors improvise, then fix the improvisation and repeat it until every detail is absorbed. At first, the actor feels stiff and mechanical. Once the actions are fully absorbed, though, the performance becomes spontaneous again (Barba 1986). The actor, therefore, is now ready to improvise.

Choreographed, but never mechanical

It is exactly this kind of preparation, I believe, that sharpens cognitive skill.

Most of the movement in Kalaripayattu is choreographed. There is a vaithari for every practice session, even for the urumi (the flexible sword). Unlike karate, there is no free sparring. In my view, the reason is the risk: the training aims at the vital spots, so even an accidental blow can do serious harm.

And yet, it is not mechanical. Two sparring partners have to be totally aligned and aware of each other's movements. When the stick does not come up in time to block a strike to the head, the striking partner has to sense it and stop the stick mid-way. With metal weapons, this becomes even more pronounced. The sequence is fixed; the attention is live. Movement scientists have a lovely phrase for this, which goes back to the Russian physiologist Nikolai Bernstein: "repetition without repetition" (Bernstein 1967).

From tacit to explicit, and back again

In Knowledge Management, we see the same journey in how knowledge moves from tacit to explicit, and then back to tacit again: the cycle that Nonaka and Takeuchi described in The Knowledge-Creating Company (1995). The master's skill is tacit. The vaithari and the sequences make it explicit. Through years of practice, the student absorbs it until it becomes tacit once more, this time as his own, ready to be reshaped into his own version.

An old Zen saying, attributed to the Chan master Qingyuan Weixin, puts this beautifully. Before study, mountains are mountains and rivers are rivers. With some learning, they cease to be. And with true understanding, mountains are mountains again, and rivers are rivers again.

There is a catch, of course. This only works if the student knows what is happening: he has to internalize the explicit knowledge and make it his own. That part cannot be outsourced.

So, where does AI fit in?

The early research gives me some pause, though it is too soon to call it settled. A study of 666 people in Switzerland found that heavy use of AI tools went hand in hand with lower critical-thinking scores, especially among younger users (Gerlich 2025). A survey of knowledge workers by Microsoft Research and Carnegie Mellon found that the more people trusted the AI, the less critical thinking they reported doing (Lee et al. 2025). Neither study proves that AI causes the decline. They do show a pattern worth watching.

The most telling result, for me, came from a small study at the MIT Media Lab (Kosmyna et al. 2025). Students who wrote essays with ChatGPT showed the lowest brain engagement, and struggled to recall what they had just written. But those who first wrote on their own, and only then turned to the AI, did well. Body first, weapon later. Meythari before kolthari!

The role of the educator, then, is less that of a reader of notes and more that of the gurukkal (the Kalari master): designing the sequence, calling out the vaithari, and watching closely for the moment a student is not ready.

That, in essence, is my answer. I do believe AI has a real role to play in education, going forward. But students still need to hone their own cognitive skills. Remembering and understanding, the lower rungs of Bloom's taxonomy, are now what AI does best. Examinations, and a student's preparedness, will increasingly be judged by the ability to analyze, evaluate and create. More importantly, even when using AI, students need to know when and why a decision was taken, and when it must be handed back to a human.

Like the sparring partner who stops the stick mid-air, the real skill lies in sensing when not to follow through.