Thinking Traces: How Not to Lose the Human in the Age of AI
Miskola methodology · Updated 2026
What are Thinking Traces in AI education? Thinking Traces is Miskola’s method for making AI-assisted student work transparent. Instead of grading only the final product, teachers see three stages — intent, direction and verification — that reveal whether a student directed the AI or simply copied its output. The method operationalizes the shift from ‘Process over Product’ in the classroom. Developed by Gatis Šeršņevs (Miskola, Latvia).
We are moving from grading results to architecting the process. Why the “final assignment” no longer proves anything, and what “text-mastery” looks like in practice?
Imagine a classroom situation: a student hands in an excellent essay. The grammar is flawless, the arguments logical, the structure impeccable. Three years ago a teacher would not have hesitated to give it a top mark. Today? Today the teacher looks at this work with suspicion. It is a “black box.” We do not know whether this is a triumph of human intellect or the result of a three-second prompt.
For centuries, traditional education focused on the Product. But in the era of generative artificial intelligence (AI), a product detached from its author has zero value, because it can be replicated infinitely. The only truly human, assessable value has shifted to the Process.
Globally, this paradigm shift is called Process over Product, and leading universities (such as Harvard and Oxford) are already beginning to experiment with “oral defenses” of written work. But in the AI Education Concept I developed (Miskola), we go one step further and offer a concrete methodology — Thinking Traces.
What are Thinking Traces?
They are not simply a saved chat history or a screenshot. They are a cognitive map that makes the “black box” transparent. To see whether a student directed the AI or merely consumed its output, we need to see three critical stages:
- Intent (The human voice): What did I want to say before I even opened the AI tool? This is the student’s original thesis. If a student cannot formulate their own idea without AI, they are not ready for the work.
- Direction (Critical selection): What prompts did I give? Even more importantly — which AI suggestion did I reject, and why? It is precisely the act of rejecting AI-generated text that shows critical thinking most vividly. This resonates with the “directive judgment” defined by Dutch researchers (Eindhoven University of Technology) — the ability to direct an algorithm rather than submit to it.
- Verification (Factual hygiene): How did I check the facts in the “synthetic content loop”? We must recognize that the internet can no longer be blindly trusted, because content there is increasingly produced by AI quoting other AI.
From theory to practice: 3 classroom scenarios
What does this look like in reality? We no longer simply ask students to “write”. We ask them to become editors, detectives and directors. Here are three ways to transform assignments according to the concept.
1. The history detective: fighting the “synthetic loop”
Students often treat AI answers as unquestionable truth. We must teach them to be skeptical.
Instead of asking them to write a report on the 1991 Barricades, we assign an AI audit. The student asks the AI to generate a description and then hunts for “hallucinations” or tonal errors in it.
A student’s “Thinking Trace” (example): “The AI wrote that during the Barricades there were widespread street battles. I deleted that sentence. After talking to my grandfather and researching ‘Barikadopedija.lv,’ I understood that the AI was dramatizing. The strength of the Barricades was nonviolent resistance, not weapons. The AI failed to grasp this cultural nuance.”
2. Literary montage: soul versus algorithm
AI texts are smooth, but often soulless and full of clichés. We can use AI as a “cognitive mirror” — to see how not to write, and then improve on it.
The task is to ask the AI to generate an epiphany in the style of Imants Ziedonis, and then perform a Style Injection. The student must rewrite the text, discarding the banalities and introducing the “Ziedonis paradox” that an algorithm cannot imitate.
A student’s “Thinking Trace” (example): “The AI offered the phrase ‘The road wound like a beautiful ribbon.’ That is far too banal. Ziedonis would never say ‘beautiful ribbon.’ I completely rewrote it as: ‘The road led nowhere. The road itself was the walker.’ I wanted to achieve that feeling of the object becoming the subject, which the AI could not grasp.”
3. Reverse engineering: AI as the student
In physics and mathematics, students often use tools to get the final answer. We can reverse the roles — let the student teach the AI (the Teaching Machine method).
The student enters a problem about Newton’s laws but deliberately omits a critical variable (for example, friction). The AI will try to solve it by making assumptions. The student’s task is to “catch” the AI in a faulty assumption.
A student’s “Thinking Trace” (example): “I deliberately did not mention that the surface is rough. The AI assumed ideal conditions and calculated an impossible acceleration. I stopped the generation with the question: ‘But did you notice that this is happening in a real environment?’ That proves I understand physics better than the tool.”
Conclusion: text-mastery as the new literacy
We are no longer teaching children to be mere “writers” in the technical sense. We are teaching them to be Directors.
I have defined this competency as text-mastery (text management). It requires the courage to stop AI generation, the courage to say “no” to a mediocre result, and the skill to inject one’s own uniquely human vision into the content.
“Thinking Traces” is not bureaucracy. It is a way for a student to say out loud: I think, therefore I am (more than an algorithm).
Practical tool: the Thinking Traces form
To make it easier to introduce this method in the classroom, I have prepared a ready-made, universal form (a reflection journal). A teacher can copy it and attach it to any writing or research assignment.
Instructions for the student: This page is more important than your final assignment. It proves that you are the director of the process. Fill it in honestly as you work.
THINKING TRACES LOG Name, Surname: _______________________________ Assignment: __________________________________ 1. STAGE: INTENT (Before opening AI) ------------------------------------ My original idea/thesis: (What exactly do I want to say in my own words, before asking for help?) __________________________________________ 2. STAGE: DIRECTION (Conversation with AI) ------------------------------------ My main prompt: __________________________________________ What did the AI offer that I REJECTED? (A specific quote or idea the AI wrote, but you deleted:) __________________________________________ Why did I reject it? (Factual error? Clichéd style? Doesn't match my idea?): __________________________________________ 3. STAGE: VERIFICATION (Fact-checking) ------------------------------------ Which fact or claim seemed suspicious to me? __________________________________________ How did I verify it outside the AI? (Book, trusted website, consultation): __________________________________________
Reference to the concept: “AI Education Concept” (Gatis Šeršņevs, 2025)
Author: Gatis Šeršņevs · Miskola (SIA Laba satura skola) · Latvia · Updated: September 2026