01000001 01100100 01110010 01101001 01100001 01101110 00100000 00101000 01000001
01100100 01101001 00101001 00100000 01010011 01010100 01000001 01001110 00100000
__ __ _____ _____ _____ ___ ____ _ ___ _____
| \/ | | ____| |_ _| | ____| / _ \ | _ \ / \ |_ _| |_ _|
| |\/| | | _| | | | _| | | | | | |_) | / _ \ | | | |
| | | | | |___ | | | |___ | |_| | | _ < / ___ \ | | | |
|_| |_| |_____| |_| |_____| \___/ |_| \_\ /_/ \_\ |___| |_|
_____ ___ ____ _____ ____ _ _ ____ _ _____ ___ ____ ____
| ___/ _ \| _ \ | ____| _ \| | | |/ ___| / \|_ _/ _ \| _ \/ ___|
| |_ | | | | |_) | | _| | | | | | | | | / _ \ | || | | | |_) \___ \
| _|| |_| | _ < | |___| |_| | |_| | |___ / ___ \| || |_| | _ < ___) |
|_| \___/|_| \_\ |_____|____/ \___/ \____/_/ \_\_| \___/|_| \_\____/
01010010 01001111 01000111 01001111 00101100 00100000 01000101 01010010 01000111
01001111 00100000 01000101 01001101 01000101 01010010 01000111 01001111 00101110

Adrian Stan
METEORAIT
FOR EDUCATORS
Pitesti, 2026
ISBN [ to follow ]
Copyright © 2026 Adrian Stan

INTRODUCTION
If you are a teacher and have felt in the last 1-2 years that
students' assignments are perhaps too perfect, that homework grades
don't match what they can do in class, that you struggle to detect who
has copied and who hasn't, and that anti-plagiarism systems can't find
anything because the text is "original" - then this guide is for you.
It's not for principals who want to "implement AI in school". It's
not for the Ministry of Education that writes strategies. It's not for
parents who want to understand "what's happening to their children".
It's for you, the Teacher who maybe enters tomorrow at 8:00 in a class
with 25-30 students, and each one has a device in their pocket
capable of generating a perfect homework in 30 seconds.
You don't have to be a computer science teacher. You don't
have to understand how ChatGPT, Gemini, Claude, or any other AI
works. You just have to recognize the problem: you can no longer
assess what the student knows if everything you assess is done at
home, unsupervised, with unlimited access to AI.
WHY NOW? Between January 2023 (the launch of ChatGPT)
and January 2026, something unprecedented in the history of
education happened: for the first time, the average student has free
access to a tool that can do all their intellectual work, instantly, at an
expert level. It's not Google (which gives you raw information that you
have to process). It's not Wikipedia (which gives you knowledge that
you have to synthesize). It's a tool that synthesizes for you, writes for
you, solves for you, thinks for you - if you let it.
The education system hasn't adapted. We still assign
homework, we still grade final "products". We still assume that the
student turns in a perfect paper because they learned something. But
the reality is that most students have discovered that they can finish
in 5 minutes what would take them 2 hours, without getting caught,
without being punished, with the same grade or even a higher grade.
It's not their fault. It's rational economics. Why wait 2 hours
when you can finish in 5 minutes? Why risk making mistakes when AI
guarantees your accuracy? Why struggle when the system doesn't
differentiate between your work and the machine's work?
The answer is not to ban AI. You can't. It's on their phone, it's
free, it's a click away. The answer is to change What and how do
you measure.
METEORAIT FOR EDUCATORS 1

WHAT YOU WILL FIND IN THIS GUIDE
Part I - Diagnosis: How to recognize cognitive atrophy in
students (FCPT - False Cognitive Power Transfer) and how to check
if you yourself have become dependent on AI for tasks you used to
do yourself.
Part II - The Weapon: Concrete protocols, logically tested, to
make AI irrelevant or counter-productive in assessment. Not
pedagogical philosophy, but practical tools: how to design a task that
cannot be solved by AI, how to check in 30 seconds if the student
understands what he wrote, how to evaluate the process (not just the
product), how to respond when parents complain or when the principal
pressures you.
Part III - Toolkit: Checklists, templates, email suggestions,
evaluation grids, classroom posters - everything you can use directly,
without major changes, from tomorrow.
Part IV - Epilogue: About limits, about giving in, about what
you can control or you can't, about why it's worth it even if it's hard.
WHAT YOU WILL NOT FIND IN THIS GUIDE
You won't find a magic solution that will transform the system
overnight. You won't find the promise that "if you do exactly this, all
students will become geniuses."
You won't find institutional validation (the ministry hasn't
approved the book, the Administration doesn't officially recommend it,
your colleagues will probably look at you strangely).
You won't find any optimistic 'pro-AI' approach that tells you it
is the future and that you must teach students to become 'Prompt
Engineers'. You will find reality: you can’t be good when use AI you if
don’t have foundation without AI. Amplification comes after
competence, not in place. AI can increase creativity if it is used for
divergence (brainstorming), not convergence (final solution).
You won't find protocols for every discipline / age or every
educational system in the world. You will find universal principles
applied on two broad axes (Math & Sciences and Humanities / 10-14
years and 14-18 years). If you are a music or chemistry teacher and
you don't find yourself in the examples, you will have to extrapolate.
METEORAIT FOR EDUCATORS 2

HOW TO READ THIS GUIDE
If you have limited time (and you probably do), read in this
order:
1. Part I, Chapter 1 (FCPT Checklist) - 10 minutes
Quickly identify if the problem exists in your students.
2. Part II, Chapter 3.2 (Micro-verification) - 10 minutes
The most effective protocol (2-3 min out of 50 minutes of class).
3. Part III (Toolkit) - 15 minutes
The templates you need for next week.
If you have time to read everything (1-2 hours), read linearly.
The guide is logically structured: from diagnosis, to solutions, to
resources, to the realism of limits. If you feel like you're giving up after
2-3 weeks of application (and maybe that's normal, fatigue is human),
go straight to Chapter 3.6 (Self-Sabotage) - the recovery protocol.
If you want to see the AI era through the eyes of Generation Z
and Alpha, read METEORAIT FOR TEENS. If you want to understand
the impact of AI on human society, read METEORAIT.
METEORAIT CONCEPTS IN BRIEF
Terms used in this guide:
1. FCPT (False Cognitive Power Transfer): The student's illusion
that he is smart because the machine gave him a smart answer. It is
the "speed intoxication" of the brain.
2. Tb (“Biological Thinking Time”): The time it takes for neurons to
make connections. It's slow, cumbersome, and "sweaty." The only one
that produces real learning.
3. Tg (“Generated Thinking Time”): The time it takes the AI to
conceive the result. It is almost instantaneous.
4. L1 (Passenger): The student who uses AI as a hammock. Risks
cognitive atrophy and irrelevance within 24 months. Measured.
5. L2 (Operator): The student who uses AI for efficiency but remains
dependent on it. Does the same thing, but faster.
6. L3 (Architect): The one who understands the process. Uses AI as
an exoskeleton (empowers it), but can operate without it (has control).
7. Useful Friction: The method by which the teacher intentionally
introduces difficulties to force the student's brain to "work" (Tb > 0).
METEORAIT FOR EDUCATORS 3

THE LAST THING BEFORE YOU START
This guide won't make you popular, it won't completely protect
you from complaints, from pressure, from fatigue. But it will make you
honest. And honesty, in a system that pretends that AI doesn't exist
or that it "will fix itself”, is already a form of resistance.
You don't have to apply everything. You don't have to be
perfect. You just have to be a little more aware tomorrow than you are
today about what you're measuring when you give a grade. And if out
of the 25-30 students you have in your class, even 5-10 end up being
able to think without "crutches”, you've done more than the system
expects of you.
This guide was written by an engineer – you will probably feel
this distinct style as you go through the book (engineering thinking is
not just an accumulation of diverse knowledge, it is a structured
method of processing reality through the filters of logic, efficiency and
systems).
Recommend that your students read "METEORAIT FOR
TEENS" - the operational adaptation of the theoretical corpus
"Cognitive Divergence Theory" (CDT) for the demographic segment
most at risk of mental atrophy: generations Z and Alpha. METEORAIT
FOR TEENS offers them a set of technical and behavioral protocols
(e.g.: "Controlled Latency", "5-Minute Wall", "Reverse Engineering")
designed to reintroduce deliberate effort into the learning process. The
ultimate goal is to move from the Cartesian paradigm of processing ("I
think, therefore I am") to the interrogative paradigm of the new era:
"Rogo, Ergo Emergo" ("I ask, therefore I become").
"METEORAIT FOR EDUCATORS is a self-defense manual for
teachers. While universities write philosophical treatises, this book
tells you what to do at school tomorrow. From the point of view of
neuroscience and modern pedagogy, it is very correct." - AI Gemini
Advanced review (Google AI), January 2026
METEORAIT FOR EDUCATORS 4

PART I - DIAGNOSTICS
CHAPTER 1: How to recognize FCPT in the classroom
(False Transfer of Cognitive Power - when the student thinks he
knows, but he only witnessed)
1.1 OBSERVATIONAL CHECKLIST
Check the behaviors you frequently see in a student. More than 7-8
ticks = high risk of FCPT.
A - Fluency without substance
[ ] Responds quickly and fluently to general questions, but gets stuck
on "Why?"
[ ] The written work is stylistically impeccable, but the student cannot
explain his/her own arguments
[ ] The vocabulary from the homework never appears in spontaneous
conversations
[ ] When you interrupt him with a side question, he completely loses
the thread.
B - Dependence on external support
[ ] Can't start a task without "checking something quickly" (euphemism
for AI / Google)
[ ] Always asks for "an example" before trying it himself
[ ] On tests (without internet access / AI), performance drops
dramatically compared to homework
[ ] Visible panic when phone is unavailable
C - Lack of metacognition
[ ] Doesn't know how to say "I'm not sure" - either keeps quiet or avoids
the answer
[ ] Can't estimate how long a task will take ("5 minutes" for everything)
[ ] Doesn't identify errors in his own work when you show it to him
after 2 days
[ ] When something is wrong, he doesn't understand WHY it's wrong
(even if you show him)
D - Asymmetric performance
[ ] Homework: 9-10 vs. Oral questions: 5-6
[ ] Written: excellent vs. Explained aloud: confusing
[ ] Technical: perfect vs. Conceptual: empty, without actual content
METEORAIT FOR EDUCATORS 5

1.2 CRITICAL DIFFERENCE: "Doesn't know" vs "Knows only with
assistance"
The student who "DOESN'T KNOW" (yet):
- Acknowledge the gap: "I didn't get to that chapter";
- Try, make mistakes, ask for specific help;
- After explanation, he can reproduce logically;
- PROGNOSIS: Can recover through normal learning.
The student with FCPT (KNOWS only with assistance):
- He doesn't recognize the gap: " I know, it just doesn't come to my
mind right now.";
- Don't try - ask directly for the answer or "a hint ";
- After the explanation, reproduce mechanically or not at all;
- PROGNOSIS: Progressive atrophy if the pattern continues.
THE DECISIVE TEST (“The 48-Hour Gap ”):
- The student submits an excellent paper;
- After two days, you ask him to explain ORALLY what he wrote;
- FCPT confirmed: Can't. Says "I forgot" or "I need to review what I
wrote in the paper."
1.3 THREE CONCRETE EXAMPLES
CASE 1: MARIA (L1 - Passenger)
Observed behavior:
- Perfect essays, textbook structure, zero grammatical errors
- When questioned: "How did you come to that conclusion?" →
Silence. Stares into space.
- When asking for rewording: repeat the exact same phrases as in
the essay
- On exams: write 60% of the usual length, lower-level language.
Diagnostic:
- Severe FCPT
- AI does 80-90% of the work, or maybe even 100%
- Maria reads, copies, possibly modifies a few words, sends
- Zero internalization.
Distinguishing marks:
- "The words are not hers" - you feel this when you read
- He has no personal opinions about what he "wrote"
- If you ask "What would you like to add?", he doesn't know.
METEORAIT FOR EDUCATORS 6

CASE 2: ANDREI (L2 - Operator)
Observed behavior:
- Math homework: all exercises correct, steps clearly explained
- At the blackboard (similar problem): knows the algorithm, but
hesitates on the details
- When you give him a new version: "I didn't do that"
- It can explain WHAT it does, not WHY it works
Diagnostic:
- Moderate FCPT
- AI is a "copilot" that Andrei uses to check
- Andrei understands 60-70%, but can't go further on his own
Distinguishing marks:
- Apparent but fragile competence
- When AI makes a mistake, Andrei doesn't detect it
- For non-standard issues: blocked
CASE 3: ANDRA (L3 - Architect)
Observed behavior:
- The work has deletions, rewritten sentences, imperfect structure
- When questioned: explains in own words, sometimes makes
mistakes, corrects herself
- When you ask her "Did AI help you?": "Yes, I asked for 3
introduction options, chose one and rewrote it"
- On exams: similar performance to homework (± 10%)
Diagnostic:
- ZERO FCPT
- AI is a tool, not an author
- Elena controls the process
Distinguishing marks:
- Can he say what he rejected from what the AI gave him and
why?
- Has stylistic preferences ("I don't like the way that sounds")
- Can operate without AI, just slower
METEORAIT FOR EDUCATORS 7

CHAPTER 2: Self-assessment for the Teacher
(Are you an L3 in your field yourself?)
2.1 WHY IT MATTERS
The Inconvenient Reality: Teachers are using AI for lesson plans,
assessments, feedback. If YOU have FCPT, you can't detect FCPT in
students.
It's not your fault. You're under pressure, tired, overworked. AI
seems like a savior.
The problem: If you become dependent on AI for basic tasks,
students sense it. Your epistemic authority diminishes.
2.2 QUICK TEST
(10 questions - answer honestly, it's just a self-assessment)
A - Assistance dependency
1. Lesson plans:
[ ] A: I write them from scratch (even if it takes me time)
[ ] B: I ask AI to generate, then significantly modify
[ ] C: I ask AI to generate, use directly or modify cosmetically
2. Assessments (tests/homework):
[ ] A: I create them myself, thinking about what I want to test
[ ] B: AI generates, I select/edit the questions
[ ] C: AI generates, I just check that I'm on topic
3. Feedback for students:
[ ] A: I write personally, based on what I have observed
[ ] B: AI suggests, I customize
[ ] C: AI writes, I just adjust the tone
B - Ability to operate manually
4. If AI fails for a week, I can:
[ ] A: I'm working normally (slower, but ok)
[ ] B: I am functioning, but I feel a lot of pressure
[ ] C: I'm panicking - I've lost the "extension" of my mind
5. The last time I read a book/article about my field (without AI):
[ ] A: In the last 2 months
[ ] B: In the last 6 months
[ ] C: I don't remember anymore
METEORAIT FOR EDUCATORS 8

C - Metacognition
6. When students ask questions I don't expect:
[ ] A: I can improvise a reasonable answer
[ ] B: I say "I'm looking and I'll get back to you tomorrow" (and I really
am looking)
[ ] C: Distract or avoid the question
7. Can I explain the material without a PowerPoint presentation /
note?
[ ] A: Yes, I can do an impromptu lesson
[ ] B: Partially - general structure yes, details no
[ ] C: No - I need constant visual support
D - Depth vs Coverage
8. When I prepare a new lesson:
[ ] A: I deeply understand over 80% of the material
[ ] B: I understand enough to teach, but I can't answer advanced
questions
[ ] C: I rely on materials already made
9. Compared to 3 years ago, I feel:
[ ] A: More competent in my field
[ ] B: Equally competent
[ ] C: Rusty - no longer practical deep
E - Relationship with “error”
10. When I make a mistake in class:
[ ] A: I admit it, I explain it correctly.
[ ] B: I subtly avoid it or minimize it
[ ] C: I don't notice it until a student points it out.
2.3 INTERPRETATION
Most answers A: You are L3. You can detect and correct FCPT in
students.
Most answers B: You are L2. You are functioning. You can detect
FCPT in students if you are vigilant.
Most answers C: You have FCPT. You need to recover BEFORE you
can help students. There is no shame. It is reality.
METEORAIT FOR EDUCATORS 9

2.4 SELF-RECOVERY PROTOCOL (if you have FCPT)
Week 1-2: Partial Detox
1. 1 lesson/week FULLY prepared without AI
2. Time it: how long does it take you?
3. Notice: where are you getting stuck?
Week 3-4: Re-training
1. Choose 1 subject you "know" but haven't taught yourself
recently
2. Explain it to a colleague (or register)
3. Identify gaps - fill them manually
Months 2-3: Stabilization
1. 50% of plans without AI, 50% with AI
2. Compare quality - where does AI really help you vs. where does
it make you lazy?
3. Redo what seems incomplete or insufficient to you.
From month 4+: Conscious symbiosis with AI
1. You use AI, but with total control
2. You can tell exactly what the AI did and what you did
3. You can operate without AI, if necessary
METEORAIT FOR EDUCATORS 10

PART II: WEAPONS
CHAPTER 3: What you can do
3.1 Anti-AI theme design
(“Assignment Engineering” - Algorithm Pitfalls)
3.1.1 BASIC PRINCIPLE
Most teachers try to detect cheating AFTER the assignment is
handed in. They check with anti-plagiarism software, look for strange
patterns, ask control questions. All of this wastes energy and
generates conflict.
There is a more effective way: the theme itself can make AI
useless or even counterproductive. You don't ban AI. You make it
irrelevant.
3.1.2 METHOD 1: Specific local context
The universal concept:
The AI is trained on global, abstract, generic data. It knows
nothing about the student's immediate reality: the neighborhood they
live in, the corner store, the bus they take, the gym teacher who
swears in the local dialect. When you insert local details into the
requirement, the AI can generate text, but it will be obviously empty
because it has no access to the personalized experience.
Application to Humanities - Language/Literature:
Vulnerable classic requirement: "Compare the main character
in novel X with a contemporary public figure."
Anti-AI requirement: "Compare the main character in novel X
with [a person from your immediate environment: the saleswoman at
the store, the bus driver you take every day, the security guard at
school]. Describe a situation in which the character would react
similarly or differently to how this person reacts when [a concrete
observable situation]."
Age differences:
If the student is 10-14 years old, ask for a comparison with
someone they see daily and about whom they can say simple, visually
METEORAIT FOR EDUCATORS 11

observable things: "how they dress", "what tone they use when they
speak", "how they react when they are in a hurry".
If the student is 14-18 years old, ask for a comparison with
someone whose motivations you can intuit: "why do you think this
person chose this job", "what regrets do you think they have", "what
would it have been like if they were in a different context".
Why it works:
AI can invent a generic person. The student who goes to the
corner store actually knows the saleswoman: he knows that the lady
is about 50 years old, smokes in front of the store, acts bossy when
there is someone present but is gentle with children. A student who
writes about "a polite and professional saleswoman" - has betrayed
himself.
Application to Math & Sciences - Physics:
The classic vulnerable requirement: "Calculate the velocity of
an object falling freely from 100m."
Anti-AI requirement: "Measure the height of the apartment
building you live in [or the school, or the tallest tree in the yard] using
a method of your choice (and explain the method). Calculate how long
it would take for an object to fall from one floor (different variations for
students) to the ground. Then test it experimentally with a safe object
(ball, empty plastic bottle) - time it or film it. Compare the real time with
the theoretical calculation. Explain the difference."
Age differences:
If the student is 10-14 years old, they accept simple
measurement methods (counting floors × 3 meters, comparing with
the height of someone they know). The experiment can be done with
the assistance of an adult.
If the student is 14-18 years old, ask for more sophisticated
measurement methods (trigonometry, shadow, smartphone apps) and
deeper explanations about the cause of the difference (air resistance,
the shape of the object, even the Magnus effect - if the object rotates).
Why it works:
The AI can do the theoretical calculation perfectly. But it can't
measure the real block, it can't throw the ball, it can't explain why real
time is different based on the concrete experiment. The student has to
go out into reality.
METEORAIT FOR EDUCATORS 12

Application to Humanities - History:
The classic, vulnerable requirement: "Analyze the impact of the
French Revolution on Europe."
Anti-AI requirement: "Interview a member of your family
(parent, grandparent, uncle) about a historical event they experienced
firsthand. Describe 3 concrete things that that person values in life
(based on real conversation, not assumptions). Then analyze: if they
had lived during the French Revolution, would they have agreed with
the revolutionary ideas? Why yes or why not?"
Age differences:
If the student is 10-14 years old, the interview can be guided
by simple questions that you can suggest: "What did you like most
when you were a child?", "What makes you happy?", "What do you
think is unfair in the world?". The historical analysis can be simple:
would they have been on the side of the king or the revolutionaries?
If the student is 14-18 years old, the student constructs his own
interview questions and makes a more nuanced analysis: what he
would have supported in different phases of the Revolution (1789 vs
1793), what contradictions he would have had between values and
context, how his opinion would have evolved over time.
Why it works:
The student must access real biographical memory about someone
they know. The AI can invent a generic grandfather who "values work
and family”, but the student knows that his grandfather doesn't tolerate
trickery, and says "Back in my day..." The authenticity is palpable.
3.1.3 METHOD 2: Clichés (“Forbidden Words List”)
The universal concept:
The AI obsessively uses certain formulas, logical connectors,
sentence patterns. They are visible to a trained eye, like a “fingerprint”.
Instead of detecting them AFTER, you ban them BEFORE. At the
beginning of the year, you clearly announce: "List of banned
words/phrases. If they appear in your essay, your grade will
automatically drop. No discussion."
Standard list (you can also complete it over time):
Generic connectors that AI often uses: "In conclusion", "On the
one hand / on the other hand", "First of all, secondly", "It is important
METEORAIT FOR EDUCATORS 13

to note that", "Thus" (at the beginning of the sentence), "Indeed", "Of
course".
Grandiose formulations without substance: "Throughout
history", "Since ancient times", "In contemporary society", "In today's
modern world".
AI's favorite pleonasms: "different points of view", "varied
perspectives", "multiple approaches".
Age differences:
For 10-14 years: Shorter list (7-10 expressions), with simple
explanations: "If you write 'in conclusion', it means you haven't thought
about how to end the text. Think about how you would end a story told
to a friend." "Soft" version: ask for reformulation, don't penalize.
For 14-18 years: Complete list (10-15 expressions). “Hard”
variant: automatic penalty of -1 or -2 points per occurrence. If the
student protests “but I really think so”, you ask: “Then explain WHY
you chose exactly this formula, not another one. If you can’t explain, it
means it’s not yours.”
Psychological effect:
The student using AI will receive the text back full of these
formulas. He has two options: rewrite everything manually (in which
case he LEARNS) or try to replace them mechanically with synonyms
(in which case the text becomes "weird" - and it shows). The student
who writes on his own naturally avoids them because he does not
think in pre-made formulas, but in his own ideas.
3.1.4 METHOD 3: Layered Requirements (“The Onion Method ”)
The universal concept:
AI is excellent at simple, clear, delimited tasks. It succumbs to
layered complexity, where level 2 depends on personal interpretation
of level 1, and level 3 requires a vulnerable personal position. Build
the theme in three layers: informational (AI can do this), interpretive
(AI can do it generically), personal-vulnerable (AI can't, because it has
no biographical experience).
Application to Math & Sciences - Mathematics:
Layer 1 - Informational: "Summarize the Pythagorean
theorem in no more than 50 words."
METEORAIT FOR EDUCATORS 14

Layer 2 - Interpretive: "Find a practical situation in your home
where the Pythagorean theorem helps you calculate something you
can't measure directly. Explain the calculation."
Layer 3 - Personal-Vulnerable: "Have you tried the
calculation in practice? If so, what was the difference between the
calculation and the actual measurement? If not, why didn't you try
(what stopped you or what were you too lazy to do)?"
Age differences:
For ages 10-14: Layer 3 can be simpler: "Have you ever
thought about this theorem when you were looking at something
around the house? What did you think about?" Or: "If you had to
explain this theorem to your younger sibling, what object in the room
would you use as an example?"
For ages 14-18: Layer 3 requires real vulnerability: "Describe
a situation in which you needed a mathematical calculation in real life
and either you didn't know how to do it, or you did it wrong, or you
were lazy. What happened?" Or even: "Why do you think this theory
seems useless to you? (If you don't, explain when you think you will
use it in reality.)"
Why it works:
Layer 1: AI writes perfectly. The student feels “safe”. Layer 2:
AI writes reasonably, but generically. The student can accept or think
for themselves. Layer 3: Everything is decided here. AI can invent a
“situation from the student’s life”, but it will be empty of real emotion,
sensory details, human contradictions. The student who writes about
“I applied and it worked perfectly” (without any problems, no surprises)
- red flag. The student who writes "I got stuck on step 3 because I
didn't understand what that formula was for" or "I didn't have a tape
measure and couldn't measure " - is real.
Application to Humanities - History:
Layer 1 - Informational: "Summarize in 80 words the causes
of World War I."
Layer 2 - Interpretive: "Identify 2 causes that the teacher did
NOT mention in class or that the textbook treats superficially. Explain
why you think they are important."
Layer 3 - Personal-Vulnerable: "If you were an 18-year-old in
1914, in one of the countries involved, would you have volunteered?
Explain honestly (not what you 'should' feel, but what you think you
METEORAIT FOR EDUCATORS 15

would have actually felt). Or: if you are not sure, explain why you
cannot decide."
Age differences:
For ages 10-14: Layer 3 can be more concrete and less
emotional: "If it were you in 1914, what would you have done when
your neighbors went to war? Would you have gone with them or
stayed home? Why?" Accept simple answers, but ask for honesty.
For ages 14-18: Layer 3 requires deeper introspection:
"Analyze what factors in your current identity (family, values, fears)
would have influenced the decision. If your answer is 'it depends on
the context,' describe 2 different contexts in which you would have
decided differently." Or: "Are you the type of person who makes
decisions based on principles or based on emotion at the time? How
do you think that would have manifested itself in 1914?"
Note for both disciplines: Layer 3 does not require "success"
or "morally correct response". It requires genuine vulnerability. "I'm
not sure what I would have done" or "I would have probably done X
out of fear, not courage" or "Honestly, I didn't try because I was lazy"
- all are valid answers. The AI doesn't know how to be vulnerable.
It will always write the "heroic" or "correct" answer.
3.1.5 THE PROGRESSION TOWARDS AUTONOMY Ramp ")
You cannot transform a "Passenger" (L1) into an "Architect" (L3)
overnight. The effort must be measured so as not to generate abandon
(panic). This progression must be done over the course of a semester:
Phase 1: Critical Consumption (L1 → L2) – “Error Detective”
Task: The student asks the AI to solve the assignment, but its
task is to find 3 weak points or an error in the answer.
Purpose: Moving from "blind acceptance" to "active verification."
Phase 2: Human structure + AI filler (L2) – “The Skeleton”
Task: The student writes the structure (content/logical steps) by
hand and asks the AI to fill in the sections. At the end, they must justify
why they chose that structure.
Goal: Regain control over the architecture of thought.
Phase 3: Sovereign Symbiosis (L3) – “The Exoskeleton”
Task: The student does all the work manually, then uses the AI
ONLY as a cognitive partner (“What did I miss?”, “Contradict me”).
METEORAIT FOR EDUCATORS 16

Purpose: AI becomes a refinement tool, not a production one.
Tip: Record progress, not just the result. A student who moved up
from Phase 1 to Phase 2 learned more than one who got stuck in AI
perfection at Phase 1.
3.2: MICRO-VERIFICATION (30 second check)
3.2.1 REALITY
A teacher with 25 students and 50 minutes of class time cannot
conduct 10-minute interviews with each student to check if they have
FCPT. The math is simple: 25 students × 10 minutes = 250 minutes =
5 hours. Impossible.
Most anti-FCPT verification systems fail because they require
time that teachers don't have. The result: teachers give up on
verification, students feel there are no consequences, the atrophy
continues undetected.
The solution is not to ask the teacher for more time, but to
change the economy of checking. Instead of checking everyone in
depth, check a few in public. The rest of the class becomes a
witness. The effect is not individual, but systemic. Note: for extremely
shy students, Micro-checking can be done at the teacher's desk, while the
rest of the class is working, so as not to block them emotionally.
3.2.2 BASIC MECHANICS
Every 2-3 classes, you choose 2-3 students. You don't
announce in advance and you don't follow a pattern. Each student gets
a single question, related to the topic they taught. You don't ask for
a summary. You don't ask for an explanation of the entire topic. You
ask for a micro detail, a decision point, a specific "why".
Timer: 30 seconds max per student. If the student answers
in 10-15 seconds, that's great. If they stutter for 30 seconds and get
nowhere, stop. Don't give public feedback at that point. Just: "Okay,
thanks." Next. Total time spent: 3-4 minutes out of 50.
3.2.3 TYPES OF QUESTIONS - MATH & SCIENCES
Principle for mathematics/physics/chemistry/computer science:
You don't ask "What is formula X?" or "What is theorem Y?"
The AI writes that perfectly and the student memorizes it mechanically.
METEORAIT FOR EDUCATORS 17

Instead, you ask about the decision process or a point of friction
that only someone who has thought through the problem could see.
Example 1 - Mathematics (equations of degree 2):
The topic was: "Solve the system of equations: 2x+3y=12 and x-y=1".
The classic (weak) question: "How much is x?"
Student: "x=3" (read from homework, memorized).
Micro-verification question (strong): "Why did you choose the
substitution method and not the reduction method? Or vice-versa, if
you used reduction, why not substitution?"
Why it works:
The student who thought through the problem knows: "I tried
reduction, but it seemed more complicated with the coefficients, so I
switched to substitution." Or: "I used reduction because x had a
coefficient of 1 in the second equation, it was simpler." The answer
has friction - it shows where he thought, where he compared options.
Student with FCPT: "I... I used what he gave me..." (red flag) or "I used
substitution because it's easier" (generic, without specific details).
Example 2 - Physics (free fall):
The topic was: "Calculate the time it takes for an object to fall from a
height of 50m."
Micro-check question: "Why did you ignore air resistance in your
calculation? Or, if you mentioned it, why did you consider it
negligible?"
For 10-14 years old: The question can be simpler: "When you threw
the ball in the schoolyard, did it fall exactly at the time you calculated
or was it different? If it was different, why do you think?"
For 14-18 years old: The question can be more technical: "If you had
taken air resistance into account, what parameters would you have
needed to know about the object? And how would you have
measured/estimated the coefficient of resistance?"
Why it works:
The student who did the experiment knows: "It was slower, I think
because of the wind" or "I couldn't time it exactly." The answer has
texture - sensory details, improvisation.
METEORAIT FOR EDUCATORS 18

Student with FCPT: "It was exactly the calculated time" (impossible in
reality) or "I haven't thrown yet, but I will" (avoidance).
Example 3 - Computer Science (code):
The assignment was: "Write a program that sorts a list of numbers."
Micro-verification question: "If you used the built -in function, why did
you choose it instead of writing the algorithm manually?"
For 10-14 years old: "When you first ran the program, did it work or
did you have an error? If you had an error, what was it?"
For 14-18 years old: "What is the complexity of the algorithm you
chose and why does this matter for a list of 10 elements versus a list
of 10,000?"
Why it works:
The student who wrote the code knows: "I had an error on line 7, I
forgot to close the parenthesis" or "I used bubble sort because the list
was small, it didn't make sense to complicate it." The answer has
history - traces of programming mistakes, of compromise.
Student with FCPT: "I chose bubble sort because it's efficient" (false,
but sounds good) or "I had no errors" (red) major flag - no one writes
perfect code the first time).
3.2.4 TYPES OF QUESTIONS - HUMANITIES
Principle for language/literature/history/social sciences:
You're not asking "What is the text about?" or "Who won battle
X?". That's raw information that the AI delivers perfectly. Instead,
you're asking about stylistic choices, personal emotions, or
connections that aren't in the text.
Example 1 - Language/Literature (essay about a character):
The topic was: "Characterize the main character in the novel."
The classic (weak) question: "What traits does the character have?"
Student: "He is brave, intelligent, loyal" (copy-paste from the
assignment).
Micro-checking question (strong): "You wrote in your essay that the
character is 'brave but impulsive '. Why did you use 'but' and not 'and'?
What did you want to emphasize by contrast?"
METEORAIT FOR EDUCATORS 19

For ages 10-14: "You wrote that the character is 'brave but impulsive'.
If you had changed the order - 'impulsive but brave' - would it have
sounded different? Better or worse?"
For 14-18 years old: "You wrote 'brave but impulsive'. Is that a
conscious stylistic choice or did you put the first option that came to
mind? If it is conscious, what do you want to suggest with this order?"
Why it works:
The student who wrote thoughtfully: "I put 'but' because I want to show
that impulsivity is a problem, not a gain" or "Honestly, I didn't think
about the difference, it just came to me." The second answer is still
valid - it shows sincerity, not copy-paste.
Student with FCPT: "I used 'but' because it's grammatically correct"
(evasive) or confused silence.
Example 2 - History (historical event):
The topic was: "Explain the causes of the Revolution of 1848."
Micro-verification question: "You mentioned 3 causes. Which of them
do you think, if they had been absent, the revolution would not have
happened? Or do you think they were all necessary?"
For 10-14 years old: "Of the 3 causes you wrote, which one surprised
you the most when you read about it? Why?"
For 14-18 years old: "Beyond the causes you mentioned, do you think
the revolution was inevitable or could it have been avoided? If so, by
what means?"
Why it works:
The student who thought: "I think that without the economic crisis,
political discontent would not have exploded so violently" or "Honestly,
I don't know which is the most important, they all seem connected to
me." Both show independent thinking.
The student with FCPT: "All causes are equally important" (diplomatic
response without substance) or chooses one of the causes, without
answering the question "Why?".
Example 3 - Native Language (argumentative composition):
The topic was: "Argue for or against the use of phones in schools."
METEORAIT FOR EDUCATORS 20

Micro-check question: "You wrote 'On the other hand, phones are
distracting'. Have you ever had this situation - being distracted by your
phone when you were supposed to be studying? What happened?"
For ages 10-14: "When you wrote the part about 'distracting phones,'
did you think of something that happened to you or someone else?"
For ages 14-18: "In your argument, you didn't mention any solid
counterarguments for the opposing position. Was a deliberate choice
(to make you seem more convincing) or did you just not think of any?"
Why it works:
The student who wrote to himself: "Yes, last week, I looked at my
phone instead of finishing my math homework" or "I thought about my
classmate who always..." The answer has a biographical subtext.
Student with FCPT: "No, that didn't happen to me" (unconvincing when
he just wrote about it) or "I generalized, it's a common phenomenon"
(speaks like a book, not like a human).
3.2.5 EFFECT
The usefulness of the system doesn't actually lie just in the 3 students
you check. It lies in the other 22 who are watching and listening.
When I see a colleague stammering for 30 seconds on a
simple question, when I hear the difference between someone who
answers with " the fingerprint of life" ("I had an error on line 7") and
someone who answers sterilely ("I used the algorithm efficiently"),
everyone's brain will recalibrate.
The next assignment, all the students will think a little more.
Not because they are scared of punishment (you didn't give any marks
in public), but because they saw the visible difference between
knowing and thinking you know.
For 10-14 years:
The effect is even stronger because students this age are
extremely sensitive to the "judgment" of the group. No one wants to
be "the one who doesn't know what they wrote." The fear of looking
stupid in front of their peers is stronger than the fear of the grade.
For 14-18 years old:
Older students may be more cynical ("I'll manage anyway"),
but they are also more reputation-conscious. A student who stutters to
a simple question about their own homework - loses their "group
esteem". Peer pressure thus becomes an ally.
METEORAIT FOR EDUCATORS 21

3.2.6 TIME SAVING (L2 Teacher Realism)
Let's be honest: most teachers are not L3. They are exhausted
L2 (Operators), with 5-6 classes, maybe over 150 students, 30-40
hours a week at school, teacher councils, parent meetings,
bureaucracy. Any system that demands "extra dedication" will fail.
Micro-checking works because it doesn't require extra
effort. You integrate it into the normal flow of the lesson: "Okay, you've
handed in your homework. Before continue, I want to clarify a few
things. Alex, you mentioned in your homework that... [question 30
sec]. Okay. Ana, you wrote that... [question 30 sec]. Okay. Paul, you
used... [question 30 sec]. Perfect. Open your book to page..."
Total: 2-3 minutes. You don't even feel the interruption. But the
students do. They feel it in their stomachs.
Realistic frequency:
You don't have to do this every hour. It's not even optimal - it
becomes predictable. Do it randomly: once every 2-3 classes,
sometimes 2 classes in a row, then a week off. Random = maximize
productive anxiety. Everyone needs to be ready at all times, because
they never know when.
The notation (or lack thereof):
You don't give grades during Micro-Checking. You just make
mental notes. If a student fails 2-3 times in a row, then have a private
conversation: "I noticed you get stuck when I ask you about your
homework. Do you want to talk?"
It's not an accusation. It's a concern. And that's even stronger.
3.3: “FLIPPED CLASSROOM 2.0”
(Radical reversal - at home AI "free", at school pure assessment)
3.2.1 WHY THE CLASSIC MODEL IS NO LONGER FUNCTIONAL
The classic model: The teacher taught in the classroom, the
student did homework at home, the teacher checked the homework.
This cycle worked for 200 years because, at home, the student was
alone with his brain, his book, and the friction necessary to learn.
METEORAIT FOR EDUCATORS 22

Now, at home, the student is no longer alone. He is with the AI.
And the AI is not like an older brother who maybe gives you a hint. It
is like a private teacher who does all your homework in 2-3 minutes, if
you ask him nicely. Or badly. It doesn't matter.
The brutal reality: Homework no longer measures what a
student can do. It measures how well a student can formulate a prompt
and how much they care about not getting caught. Almost all students
take the shortcut. Not out of moral laziness, but out of rational
economy: why sit for 2 hours when you can finish in 5 minutes and get
the same grade?
Teachers are struggling to detect cheating, trying to invent
"anti-AI" homework, asking control questions. It's an arms race that
you might lose. Because AI is evolving faster than your detection
methods. And because you don't have time to check 150 homework a
week with the attention of an investigator.
The solution is not to fight reality. The solution is to accept
reality and use it to your advantage.
3.3.2 THE NEW PARADIGM (Radical reversal)
Give up the idea that you can know what the student is doing
at home. You can't. There's no point in trying anymore. Instead,
redraw what's happening at home and what's happening at school.
HOME = Free preparation (AI is allowed, even encouraged):
The student can use AI for whatever they want: to generate
ideas, to structure, to ask for explanations, to verify information, to
compare options. At home, it's the chaotic laboratory where there are
no rules. AI is the brainstorming partner, the research assistant, the
hypothesis generator.
Important: You are not asking the student to submit anything
completely written at home. You are only asking for preparation:
outlines, personal notes, questions they have asked themselves,
options they have considered and rejected. Raw materials, not
finished products.
AT SCHOOL = Monitored execution (zero AI, zero phone):
This is where the things that matter happen: handwriting, on-
the-spot synthesis, oral explanation, real-time problem solving under
observation. Everything that is assessed happens in the classroom,
right in front of you, without access to any devices.
METEORAIT FOR EDUCATORS 23

It's not a continuous exam. It's real work. The difference is that
the work happens where you can see it, not in some invisible place
where the AI can do 80% of the work without you knowing.
Why it works:
Because you're no longer fighting against the current. You let
the AI do what it does well (generation, volume, speed) at home,
where you have no control anyway. And you focus the evaluation on
what the AI can't do: thinking under pressure, personal synthesis,
visible cognitive friction.
Bonus: The student no longer feels like you're tricking or
trapping them. The rules are clear, transparent, honest. "At home, use
whatever you want. At school, just you and your brain."
3.3.3 PRACTICAL IMPLEMENTATION - MATH & SCIENCES
Example 1 - Mathematics (equations):
The classic (dead) model:
Home: Solve 10 equations.
At school: Check the solutions, move on to the new chapter.
Problem: 8 out of 10 students copied the solutions from the AI. You
correct 250 equations, of which 200 are "fake".
Flipped 2.0:
HOME (Preparation - Free AI):
"Prepare for tomorrow. Understand how to solve equations of
type X. Use whatever resources you want (AI, YouTube, textbook,
explanations from colleagues). Come with questions if something is
not clear to you. You don't have to hand in anything written."
Or the more structured version: "Ask the AI to solve 3
equations and explain each step to you. Write down on a sheet (just
for you) what steps you didn't understand or what surprised you."
AT SCHOOL (Evaluation - zero AI):
First 20 minutes: Each student solves 3 equations on a sheet,
by hand. Hand in the sheet with their name. Different versions for each
bench (so they don't copy from their neighbor).
You correct 75 equations (25 students × 3), not 250. And you
know that they are all real, because they wrote them in front of you.
METEORAIT FOR EDUCATORS 24

For 10-14 years:
The equations can be simpler, the time can be 25-30 min instead of
20. You can allow a "warm-up" equation that they do together at the
board, then 2-3 alone.
For 14-18 years old:
The equations can be more complex or you may also ask for a written
explanation: "Choose one of the 3 equations and explain why you
chose method X, not method Y."
Example 2 - Physics (formula problems):
Flipped 2.0:
HOME:
"Tomorrow, we have a test on the application of Newton's laws.
Prepare however you want. If you want to ask the AI to generate 10
problems for you and solve them, perfect. If you want to solve them
from the textbook, perfect. If you want to watch YouTube for
explanations, perfect. The only rule: come prepared."
Plus, option: "If you find a problem that the AI solves incorrectly
or that you don't understand, write it down. We'll discuss it tomorrow."
AT SCHOOL:
First 30 min: Students have to solve 2 problems on the sheet.
One direct (formula application), one indirect (requires you to think
about which law you are applying). I hand over the sheet.
Next 15 min: Group discussion - did anyone find an AI bug?
Did anyone have a different approach? We learn from diversity, not
uniformity.
Example 3 - Computer science (programming):
Flipped 2.0:
HOME:
"Tomorrow we're writing code in class for problem X. Prepare
at home: try to understand the algorithm, the structure, the logic. If you
want to ask the AI to generate code and study it, that's fine. If you want
to imprison yourself too, and compare with what the AI generates,
even better."
AT SCHOOL:
Students write code from scratch, on paper or on a computer
(but without internet/AI access). It doesn't have to be perfect - syntax
errors are okay. You're evaluating logic, not perfection.
METEORAIT FOR EDUCATORS 25

For 10-14 years:
The code can be pseudo-code or a flowchart, not necessarily exact
syntax. What's important is the logic: "If X, then Y, otherwise Z."
For 14-18 years old:
Functional code, correct syntax. Bonus: "Explain why you chose this
data structure, not another."
3.3.4 PRACTICAL IMPLEMENTATION - HUMANITIES
Example 1 - Native Language (argumentative essay):
The classic (dead) model:
Homework: Write a 500-word essay.
At school: The teacher corrects 25 essays, of which over 20 are written
by AI, with or without "cosmetic" changes.
Flipped 2.0:
HOME (Preparation - Free AI):
"Tomorrow we are writing an essay in class about topic X.
Prepare: think of 3-5 arguments for or against. You can ask the AI to
generate 10 arguments for you and choose the best ones. You can
talk to family/friends. You can read articles. Come tomorrow with
ideas, not written text."
AT SCHOOL (Evaluation - zero AI):
First 30-40 minutes: Each student writes the essay, by hand,
on a sheet of paper. They can refer to their homework sheet of notes
(encourage them to have it), but they can't copy the entire text - it's
just a reminder note.
You correct 25 essays, handwritten, in different calligraphy,
with erasures, with personal style. You can see who is thinking and
who is repeating mechanically.
For 10-14 years:
The essay can be shorter (200-300 words), the structure can
be simpler (introduction, 2 arguments, conclusion). You can allow 5-
10 minutes of discussion in groups of 2-3 before the writing stage - to
test their ideas orally.
For 14-18 years old:
Full essay (500 words), complex structure. You can also ask
for a self-evaluation at the end: "What argument do you think is the
weakest in your essay and why did you keep it?"
METEORAIT FOR EDUCATORS 26

Example 2 - History (event analysis), Flipped 2.0:
HOME:
"Tomorrow we're writing about the causes of the Revolution in
X. Prepare yourself: read the materials, ask the AI if you want to
explain what you don't understand, compare different sources. If the
AI gives you an explanation that contradicts the textbook, write it
down, and we'll discuss it tomorrow."
AT SCHOOL:
First 30 min: Each student writes an analysis (1-2 pages) about
2-3 causes they consider essential. It doesn't have to be all the causes
from the textbook - they can choose the ones they think are most
important and argue why.
Next 10 min: Micro-check - 3 students orally explain why they
chose the respective causes.
Example 3 - Foreign language (composition), Flipped 2.0:
HOME:
"Tomorrow we're writing in [foreign language] about topic X.
Prepare: think about the vocabulary you need, the structure, the verb
tenses. You can use AI to generate example sentences for you, to
check grammar, for anything. Come prepared with clear ideas."
AT SCHOOL:
First 35 min: Writing in [foreign language], by hand. I can have
a physical dictionary on my desk (not my phone). You rate content +
grammatical correctness.
For 10-14 years:
The essay can be shorter (150-200 words), with a more concrete topic
("Describe your room" instead of "Argue about the importance of
education").
For 14-18 years old:
More complex composition (300 words), abstract topic. You can also
ask for the use of specific grammatical structures: "Use at least 2
conditional sentences and 1 relative clause."
3.3.5 PREDICTABLE RESISTANCE
When you announce Flipped 2.0, you will encounter 3 types of
resistance:
METEORAIT FOR EDUCATORS 27

1. From students: "But why should I write by hand? It's 2026!"
Answer: "Because your hand and your brain are connected in a way
that a keyboard doesn't. When you write by hand, you're forced to think
slower, to choose your words more carefully, to feel the friction. This
is training, not punishment. Just like an athlete runs on sand to be
faster on asphalt."
More direct version (for 14-18): "Because I want to see what you can
do, not what the AI can do. If that annoys you, ask yourself why."
2. From parents: "My child has a lower grade now!"
Answer: "Normal. Before, he was grading the AI's work. Now he's
grading his own work. The difference you see is not that he learned
less, but that before you had an illusion. I'd rather know the truth
now, when we can correct it, than find out in 5 years when it's too late."
Gentler tone (if the parent is concerned, not aggressive): "I understand
that it's worrying. But the real grade shows where we need to help the
child grow. The fake grade shows that he knows how to use a phone."
3. From colleagues: "I don't have time to correct in class!"
Answer: "You don't correct in class. You correct at home, like always.
The difference is that now you correct real papers, not fake papers.
You save time, you don't waste it. You don't stop wondering 'is this
written by him or ChatGPT?' - you know for sure."
3.3.6 TIME SAVING
Let's do the math:
The classic model:
- Homework: 25 students × 2 pages = 50 pages to correct
- Proofreading time: 50 pages × 3 min/page = 150 minutes (2.5
hours)
- Maybe 80-90% are generated by AI → you waste 1.5-2 hours
correcting “ fake ” text
- Plus: verification questions, suspicions, conflict with students
who deny
Flipped 2.0:
- Classroom assessment: 25 students × 1-2 pages written in 30-
40 min
METEORAIT FOR EDUCATORS 28

- Proofreading time: 25-50 pages × 3 min/page = 75-150 minutes
(1.2-2.5 hours)
- Of which: 100% are real → zero time wasted on fake text
- Zero suspicions, zero "detection" conflicts
PLUS: Invisible bonuses:
1. Students come more prepared (because they can't rely on AI
during assessment)
2. You correct faster (because texts have personality, they are not
standardized by AI)
3. Your feedback is more relevant (because you see real errors,
not artifacts of poorly edited AI)
Conclusion: You don't necessarily save time on proofreading. But
you save mental time - you don't have to doubt, check, argue. And
the quality of the information you receive is infinitely superior.
3.3.7 GRADUAL VARIANT (if you can't make a radical leap)
If you are in a rigid system or have students/parents who
cannot accept sudden change, implement progressively:
Month 1-2: Testing (announce transparently)
"For the next 2 months, we're going to experiment. Once a week,
homework will be done in class, not at home. The rest of the week, as
before. I want to see the differences."
Month 3-4: Expansion (if it works)
"I've seen it work. We're increasing to 2 out of 3 homework
assignments in class."
Month 5+: Normalization
"From now on, everything that is assessed is written in class. At home
it is for free preparation."
Or, if you can't control the system (the Administration requires X
homework over a period of time), you do a hybrid:
1) "Formal" homework (noted in the catalog) = in class
2) "Preparatory" homework (ungraded) = at home, with free AI
This way you meet bureaucratic requirements, but you evaluate
realistically.
METEORAIT FOR EDUCATORS 29

3.2.8 SOVEREIGN SCHOOL
(Data Privacy and Local AI)
In an international context (GDPR in Europe, COPPA/FERPA
in the US), the use of AI raises huge privacy issues for student data.
The problem: When students use commercial versions (ChatGPT,
Gemini, etc.), their data (and their thinking errors) are stored on
external servers.
The Solution: We recommend schools explore “Local AI.” Using free
software (such as Ollama), a school can run language models (Llama
3 or Mistral) on a dedicated computer or multiple computers in the
computer lab, without an Internet connection.
Advantage: Zero data leaks. Zero subscription costs. Total cognitive
sovereignty over the learning environment.
Disadvantage: Consistent material investment, PCs must be new
generation, with sufficient processing power and memory.
3.4: KNOWLEDGE ASSESSMENT
(Visible History + Equation 50% Tb / 50% Tg)
3.4.1 WHY CLASSIC EVALUATION IS OUTDATED
The classic evaluation measures only one thing: the final
product. The essay has 500 words, correct structure, zero
grammatical errors, coherent arguments. Note: 10. The problem is that
you don't know if the essay was written in 2 hours of cognitive sweating
or in 3 minutes of copy -paste, possibly followed by 5-10 minutes of
beautification.
The final product tells you nothing about the process. And the
process is everything. The process is where learning happens,
where neurons connect, where error becomes lesson. When you only
evaluate the product, you reward those who hide the process best.
That is, those who use AI elegantly and leave no trace.
The solution is not to become a detective. The solution is to
change what you measure.
Instead of just measuring "What did you deliver?", you
measure "How did you get there?" The process becomes visible,
measurable, valuable.
METEORAIT FOR EDUCATORS 30

3.4.2 VISIBLE HISTORY (“VERSION HISTORY”)
The basic concept:
When you write something on a computer, every change
leaves a trace. Google Docs, Microsoft Word, any modern platform
keeps a complete log: what you wrote, when you wrote it, what you
deleted, what you rewrote, how long it took between the first draft and
the final draft. This data is a direct window into the process.
A student writing alone: erasures, rewordings, pauses for
thought, gradual growth of the text. A block of 50 words appears in 5
minutes, then a 3-minute break, then another 40 words, then an
erasure, then rewritten.
A student using AI: a 300-word block appears in 30 seconds.
Then 5 minutes of superficial editing - changing "thus" to "therefore",
"in conclusion" to "in order to conclude".
Implementation in Humanities (essay/composition):
Ask students to write in Google Docs with " Version History "
enabled (it is enabled by default - no special setting is required). When
submitting the assignment, you not only receive the final document.
You also receive the " View Version History " or check it yourself.
What are you looking for in history:
Organic growth: the text grows gradually, perhaps 50-100
words per section of text, with pauses in between. A sign that the
student is thinking, writing, stopping, thinking again.
Significant deletions: the student wrote a paragraph, erased it
completely, wrote something else. This is active thinking - "no, that
doesn't sound right, I'll try something different."
Total time: a 500-word essay written by a human takes
between 30-90 minutes (depending on age and experience). If the
total time is under 15 minutes, red flag major.
For 10-14 years:
Expectations are more relaxed. A 300-word essay written in
40-60 minutes is reasonable. Lots of erasures and rewording are
normal - thinking is still forming. You can teach them: "I want to see
that you tried, that you got stuck somewhere, that you rewrote. That
shows that you're working with your brain."
METEORAIT FOR EDUCATORS 31

For 14-18 years old:
Expectations are rising. A 500-word essay written in 30-50
minutes is reasonable for an experienced student. Too few erasures
can also be a red flag - means the text came too "clean" too quickly.
A good essay has scars.
What to do when you catch fraud through Version History:
You don't accuse publicly. You send the student a private
message: "I noticed in the document history that a block of 400 words
appeared in 10 seconds. That's physically impossible to type. Do you
want to discuss?" It's not an interrogation. It's a conversation.
Sometimes the student admits, sometimes he has an explanation (he
wrote in another document and copy -pasted from his own older text -
verifiable). But, more often than not, his silence confirms.
Limitation of the method:
It only works for work done on the computer. If the student
writes in another editor (Word offline, Notepad, even on paper) and
then copy -pastes into Google Docs, the history is useless. That's why
this method is combined with other methods - it's not the only one.
3.4.3 “VERSION HISTORY” ON PAPER
If the student writes by hand (as I suggested in Flipped
Classroom 2.0), you can see the process directly on paper, without
technology.
What are you looking for:
Visible erasures: words crossed out with a line, paragraphs
circled and moved, arrows restructuring the order of ideas. These are
the scars of authentic thinking.
Inconsistent handwriting: when you write quickly, then stop and
think, then resume writing, speed shows in your handwriting. A
uniformly written text, without variation in pressure or speed, is
suspicious (either copied from somewhere, or copied from a neighbor,
or rewritten "cleanly" after it was generated).
Odd spacing: If the student left large spaces between
paragraphs because they "didn't know what to write there”, then filled
them in later (with different ink or different pressure), that's noticeable.
And it's normal. It's thinking.
METEORAIT FOR EDUCATORS 32

Full margins: Thinking students often write notes in the
margins, alternative ideas, questions for themselves. If the paper is
"clean" like a printed document, it's suspicious.
For 10-14 years:
Erasures and visual chaos are very common. Don't penalize
for it - it's part of the process. You can even encourage it: "I want to
see that you've corrected yourself, that you've changed something. If
the sheet is too clean, I think you've rewritten it from a draft and maybe
lost good ideas in the process."
For 14-18 years old:
Older students may have more stable handwriting and fewer
visible erasures - they've already developed a writing style. But you
should still see traces of thought: words replaced; ideas restructured.
If the text looks like a museum manuscript - it's suspicious.
3.4.4 EQUATION 50% Tb / 50% Tg (Biological vs Generated)
Theoretical concept: Tb = Biological Time - the time the human
brain spends processing, thinking, building. Tg = Generated Time -
the time the machine spends generating output.
In a world without AI, Tg was zero. It was always Tb. Now, Tg
can now be 90% of the process, and Tb only 10% (speed reading,
approval, beautification, submission). When Tb drops below a critical
threshold, atrophy is inevitable.
Practical translation for evaluation:
Instead of giving all the grade on the final product (dominant
Tg), divide the grade into two equal components:
- 50% for the product (Tg): how good the text is, how correct
the exercise is, how coherent the argumentation is. That's what you've
always evaluated.
- 50% for process (Tb): how much time the student spent
thinking clearly, how much they struggled with ideas, how much they
made mistakes and corrected themselves. This is new.
Implementation in Math & Sciences - Mathematics:
Tg component (50% of the grade): Is the final solution, correct? Is
the numerical answer, correct? Are all formulas applied correctly?
Component Tb (50% of the grade): Did the student show the
intermediate steps? When he made an error (normal in the solving
METEORAIT FOR EDUCATORS 33

process), did he visibly correct it or did he completely erase it and
rewrite "cleanly"? Did he explain why he chose a particular method?
Concrete example:
Student A: Perfect solution, without any errors, written cleanly,
without erasures. Correct result. Tg score: 10/10 (the product is
perfect) Tb score: 3/10 (the process is invisible - either he is a genius
or he copied, probably the second option) Final score: (10+3)/2=6.5.
Student B: Solution with 2 minor errors, one method tried and
abandoned (visible on the sheet), explanation written at the end: "I
tried method X but it didn't work for reason Y, and then I used method
Z and that's how it worked." Tg grade: 7/10 (product has minor errors)
Tb grade: 9/10 (the process is completely visible; the student is
thinking) Final grade: (7+9)/2=8.
The apparent paradox: Student B, who made more mistakes,
gets a higher grade. It's not a paradox. It's recalibration. Student B
learns. Student A consumes.
For 10-14 years:
The Tb component can be more generous. If the student tried, even
if the method was completely wrong, Tb can be 6-7/10. "You thought,
you tried, it didn't work out, but you learned something from the
mistake." The product (Tg) may be weak, but the effort (Tb) is visible.
For 14-18 years old:
The Tb component becomes more demanding. It's not enough to
try, they have to show why they tried that method, what made them
abandon it, what they learned. Tb requires visible metacognition, not
just "I tried and it didn't work."
Implementation in Humanities - Essay:
Tg component (50% of the grade): Is the final essay coherent? Does
it have a logical structure? Are the arguments well supported? Is the
grammar correct?
Component Tb (50% of the grade): How much time the student
spent writing (from “ Version History ” or estimated from the complexity
and length of the text)? How many visible rewordings? Did he reject
ideas and explain why?
Concrete example:
Student C: Perfect essay, 500 words, zero mistakes, written in
12 minutes according to Google Docs History, with a 300-word block
METEORAIT FOR EDUCATORS 34

appearing in 10 seconds. Tg score: 10/10 (the product is perfect) Tb
score: 2/10 (the process is almost absent) Final score: (10+2)/2=6.
Student D: Good but not perfect essay, 500 words, a few minor
mistakes, written in 55 minutes, with 4-5 visible rewords, one
paragraph completely deleted and rewritten differently. Tg score:
7.5/10 (product has small imperfections) Tb score: 9/10 (process is
very visible, authentic thinking) Final score: (7.5+9)/2=8.25.
Student D thus gets a better grade. Not because the product is
better, but because the process is real.
3.4.5 CONCRETE EVALUATION GRADE
GRID FOR REAL SCIENCES (Mathematics/Physics/Chemistry):
COMPONENT Tg - Final product (50 points):
Correctness: Is the final answer, correct? (0-20 points)
Method: Is the formula/algorithm applying the right one? (0-15 points)
Presentation: Are the steps clearly written, not just the final answer?
(0-15 points)
COMPONENT Tb - Visible process (50 points):
Multiple attempts: Are alternative methods seen tried? (0-15 points)
[ ] 0 points: No visible attempt, everything "clean"
[ ] 7-8 points: A visible attempt, then the correct method
[ ] 13-15 points: Multiple trials, comparisons between methods
Corrections: Errors seen and visibly corrected? (0-15 points)
[ ] 0 points: No visible errors (suspect) or uncorrected errors
[ ] 7-8 points: Some minor corrections
[ ] 13-15 points: Major corrections, shows real-time learning
Explanation: The student explains WHY he/she chose a method? (0-
20 points)
[ ] 0 points: No explanation
[ ] 10 points: Minimal explanation ("I used formula X")
[ ] 18-20 points: Full explanation ("I tried Y but it didn't work because
Z, I switched to X because...")
For 10-14 years: Lower the expectations on "Explanation" - 10 points
for minimal explanation is enough. Increase the generosity on
"Multiple Attempts" - any visible attempt gets points, even if the
method is completely wrong.
METEORAIT FOR EDUCATORS 35

For 14-18 years old: Raise expectations - explanations need to be
more sophisticated, showing conceptual understanding, not just
mechanical application.
GRID FOR HUMANITIES (Native Language / History)
COMPONENT Tg - Final product (50 points):
Content: Are the arguments solid and relevant? (0-20 points)
Structure: Does the text have logical coherence? (0-15 points)
Language: Grammar, spelling, appropriate vocabulary? (0-15 points)
COMPONENT Tb - Visible process (50 points):
Time invested: How much time did you spend writing? (0-15 points)
[ ] 0 points: Under 15 minutes for 500 words (naturally impossible)
[ ] 7-8 points: 15-30 minutes (fast, but possible)
[ ] 13-15 points: 30-60 minutes (reasonable for thinking + writing)
Reformulations: How many ideas were rewritten/restructured? (0-15
points)
[ ] 0 points: "Clean" text, no changes (suspicious)
[ ] 7-8 points: Some minor changes (words replaced)
[ ] 13-15 points: Major rewording (rewritten paragraphs, moved ideas)
Vulnerability: Does the text have a personal voice or is it sterile? (0-20
points)
[ ] 0 points: Generic text, sounds like a textbook
[ ] 10 points: A few moments of personal voice
[ ] 18-20 points: Distinctive voice, personal opinions, examples from
own experience
For 10-14 years: "Vulnerability" can be assessed more leniently - it's
enough to have a personal phrase ("I think..." or "when I saw..."). For
"Time invested", 40-60 minutes for 300 words is quite reasonable.
For 14-18 years: "Vulnerability" requires more depth - not just "I
believe", but reasoning why you believe so, expressed doubts,
acknowledged contradictions. On "Time invested", under 30 minutes
for 500 words becomes red flag.
3.4.6 HOW TO COMMUNICATE THE NEW EVALUATION
When you announce that from now on 50% of the grade is on
the process, not the product, you will inevitably hear:
METEORAIT FOR EDUCATORS 36

From students: "But this is unfair! I worked hard and didn't make
any mistakes, why should I get a low grade?"
Answer: "Actually, it's the other way around. If you worked hard
and it shows, you get a higher grade even if you made a mistake. The
new assessment protects you, not those who copy elegantly. If you
made a mistake, and you see 10 attempts on the sheet, you get an 8.
If you answered perfectly, but the sheet is suspiciously clean, you get
a 6. It's your choice."
From parents: "My child did everything right, why didn't he get
the maximum grade?"
Answer: "The product is perfect, that's true. But the process is
invisible. Either your child is a genius who never makes mistakes (in
which case, congratulations, but it's statistically almost impossible), or
someone else did the hard work. I'm not blaming anyone. The new
assessment simply demands transparency. If the child shows how
they arrived at the answer next time, the grade will reflect that."
From colleagues: "That's subjective! How do you measure the
'process'?"
Answer: "It's less subjective than you think." Version " History
is objective - time is time. The erasures on the page are objective -
either they are or they are not. And yes, the 'vulnerability' or
'explanation' component has a dose of subjectivity, just like grading an
essay 'good' vs. 'excellent'. But at least now we're measuring
something that matters - learning - not just the ability to hide the
source."
3.4.7 BONUS: SELF-EVALUATION (Student self-evaluation)
An advanced option, especially for 14-18 years old: ask the
student to evaluate their own process before submitting the
assignment.
At the end of the assignment, add a mandatory section: "Self-
assessment (does not influence the grade, but if missing, it means -1
point for incomplete submission)".
Questions: How long did you spend working on this? (honest
estimate) What was the hardest part? Why? Did you use AI? If so, for
what? (generation, verification, explanations, other) If you had to redo
METEORAIT FOR EDUCATORS 37

the assignment without any external help, what percentage of it could
you reproduce (0-100%)?
The last question is crucial. If the student honestly answers
"30%", you know they have FCPT. If they answer "90%", then they fail
the Micro-check, you know they are lying or don't know their own limits
(both indicate problems).
Self-assessment doesn't directly change the grade, but it gives
you context. And it forces the student to reflect: "Can I really reproduce
this myself?"
3.5: Predictable resistance (Students, Parents, Education System)
3.5.1 WHY RESISTANCE IS INEVITABLE
When you change the rules of the game, players who were
winning under the old rules (maybe) will resist. Not out of malice, but
out of survival instinct. Students who were getting high grades with
minimal effort will feel threatened. Parents who saw "success" in their
child may feel uncomfortable when their grades drop. Colleagues who
have been doing the same routines for 20 years will feel attacked.
Resistance doesn't mean you're wrong. It means you've hit a
nerve. If you implement the METEORAIT protocols and no one objects
- it means you haven't really implemented them. Resistance is a sign
that the old system is feeling the pressure.
The point of this chapter is not how to avoid resistance. It's how
to be prepared for it. To have clear, firm answers, based on data, not
emotion. To transform the teacher from "the one who complicates
things" to "the guardian of mental health."
3.5.2 STUDENT RESISTANCE
WHAT YOU WILL HEAR (Top 5 objections):
1. "But all the other teachers let us use AI! Why are you different?"
2. "It's 2026! Why write by hand when there's technology?"
3. "It's not my time! If AI can do this in 5 minutes, why should I wait 2
hours?"
4. "But I really think so! I didn't copy, my style resembles AI!"
5. "I'm going to do badly in the exams! The others have high grades
with AI, I have low grades without!"
METEORAIT FOR EDUCATORS 38

HOW TO ANSWER:
Response to objection 1 - "The other teachers allow it":
Version for 10-14 years (milder): "Other teachers may have
different rules. Each teacher decides how they evaluate. I chose to
evaluate what you can do, not what the machine can do. If this makes
you uncomfortable, that's normal - it means you're getting used to
being helped too much. But discomfort is exactly where you grow."
The 14-18 version (more direct): "Other teachers have other
priorities. I prioritize your real ability. If you want high grades without
effort, this is the wrong class. If you want to get to college or a job and
not be outdone by those who really know what they're doing, stay here
and we'll work together. You choose."
The authoritative version (for both ages): "It's not a personal
rule. It's based on research on cognitive atrophy. When you constantly
delegate thinking, your brain adapts to not thinking anymore. It's called
FCPT - False Cognitive Power Transfer. I don't make up rules, I
enforce mental safety protocols."
Answer to objection 2 - "Why write by hand?":
"Because the hand and the brain are directly connected. When
you write by hand, you think slower, you choose your words more
carefully, you feel the friction. This is like a workout, like athletes who
run with weights do - not because there is no electric bike, but because
they want to be stronger. Do you want to be stronger or more
comfortable?"
With authority: "There are neurological studies that show that
handwriting activates areas of the brain that typing doesn't. It's not
nostalgia, it's biology. If you want links, I'll give you some."
Response to objection 3 - "It's not my time!":
"That's right, it's not your time. But learning isn't about time
either, it's about transformation. If AI does your homework in 5
minutes, you save 1 hour and 55 minutes. The question is: what do
you do with that time? If you invest it in learning more deeply, perfectly.
If you spend it scrolling on TikTok or Instagram, you've wasted not only
time, but also the opportunity to become more capable."
Harder version (14-18 years old): "Your time is no more
valuable than your brain. If you choose to save time now and become
METEORAIT FOR EDUCATORS 39

addicted to machines, in 10 years you will pay with years of
professional stagnation. The decision is yours, just as the future is
yours."
Response to objection 4 - "My style resembles AI!":
"It's possible. Let's test it: take your last assignment and
explain to me verbally, without looking at it, what the main idea was
and why you structured it that way. If you can, great - the style is yours.
If you can't, then it wasn't your style, it was the machine's. It's not an
accusation, it's a diagnosis."
Plus: "If your natural style resembles AI, it means you think in
clichés and generic formulas. That's not a quality, it's a signal that you
need to develop your personal voice. That's exactly why we do these
exercises."
Answer to objection 5 - "I'll do badly in the exams!":
"False. You'll be the only one prepared. They're getting high
grades now because the AI is doing their homework. When the exam
comes, when the AI disappears, they're left with nothing. You've been
thinking for weeks on your own - when the exam comes, you're already
trained. They're like a bodybuilder on steroids taking pictures on
Instagram. You're like a wrestler training in an empty gym. Who do
you think is going to come out on top when the real match comes?"
The milder version (ages 10-14): "It's the other way around.
The others seem good now because the machine helps them. But
when the big test comes (exam, interview, real-life situation), the
machine is no longer there. You will be prepared, they won't."
3.5.3 PARENTAL RESISTANCE
WHAT YOU WILL HEAR (Top 5 objections):
1. "My child has lower grades now! What did you do?"
2. "You're preventing him from using technology! You're preparing him
for the past, not the future!"
3. "The other kids have higher grades with AI. Why is my child at a
disadvantage?"
4. "You have no right to ban AI! It's discrimination!"
5. "My child says you are too strict. He feels stressed."
METEORAIT FOR EDUCATORS 40

HOW TO RESPOND (Email/face-to-face conversations):
Response to Objection 1 - "Lower Grades"
Email (formal):
"Dear [Parent Name],
I understand your concern about [Student Name]'s grades. Let
me clarify what has changed and why.
Before: [Student Name] was handing in assignments that were
graded and the grade reflected the quality of those assignments.
Problem: I wasn't sure how much of those assignments were the
child's work and how much was the AI assistant's work.
Now: I assess what your child can produce independently,
under observation, in the classroom. The grade reflects actual ability,
not assisted ability.
The difference you see doesn't mean [Student Name] learned
less. It means you're seeing the truth now. I'd rather know the truth
now, when we can help, than discover in 5 years on a major exam or
job interview that the foundation is fragile.
Reference: According to studies on Cognitive Atrophy through
Delegation (a phenomenon documented in educational psychology),
constant externalization of thinking leads to a decrease in the capacity
for independent processing. This is not personal opinion, it is a
documented mechanism.
Proposal: Let's discuss together how we can help [Student
Name] develop his or her true potential.
Face-to-face conversation (more direct):
"I understand that it's worrying. But let's be honest: would you
rather have me see now that [Student Name] has gaps or would you
rather discover in 3 years when you take the Final Exams? The old
grade was a comfortable illusion. The new grade is the uncomfortable
reality. But reality can be corrected. Illusion cannot."
Response to Objection 2 - "Hinder Technology"
"I'm not stopping him from using technology. I'm teaching him
to use it without being dependent on it. The difference is crucial.
Your child will live in a world full of AI. The question is not whether he
will use AI, but whether he will control AI or be controlled by AI."
METEORAIT FOR EDUCATORS 41

"According to Cognitive Divergence Theory, the population will
stratify into 3 levels: L1 (Passengers - completely dependent on AI),
L2 (Operators - operate AI but do not deeply understand), L3
(Architects - control the process). In 10-20 years, the difference
between these levels could reach tens of times differences, in terms
of capacity. I am preparing [Student Name] to be L3, not L1."
Response to Objection 3 - "Others have higher grades"
"It's true. The others have higher grades because I can't control
what other teachers do. But think about it this way: if 10 teachers give
swimming certificates, perhaps without checking if those kids can
swim, and I'm the only one who actually puts them in the water,
naturally my grades seem lower. But when it comes to real competition
(national exam, college, life), who do you think will cope?"
Diplomatic version: "I don't judge other students - everyone has
their own methods. But I choose to evaluate what your child can do on
their own. If that means temporarily lower grades, I accept it. Because
I'd rather [Student Name] have a solid foundation than have a high
GPA on quicksand."
Response to Objection 4 - "Discrimination"
"I don't ban AI. I allow it freely at home for preparation. What I
don't allow is to evaluate the AI's work instead of the child's work. It's
not discrimination, it's honesty. If [Student Name] uses AI at home to
learn (understand concepts, check logic), great. If he uses it to avoid
thinking, then the grade reflects that. I don't punish the use of AI, I
punish the absence of his own thinking."
Legal plus (if the parent threatens): "The school regulations
allow me to establish the evaluation methods as long as they are
transparent and applied consistently. I have clearly announced the
rules from the beginning. All students have the same rules. I do not
discriminate against anyone - everyone is evaluated according to the
same criteria: the ability to produce independently."
Response to objection 5 - "The child is stressed"
The empathetic version (for worried, non-aggressive parents):
"I get it. Change is stressful. [Student's name] was used to a certain
way of working and now he has to adapt. But the stress he feels now
is productive stress - it's a signal that his brain is working, adapting,
METEORAIT FOR EDUCATORS 42

growing. It's like at the gym: muscles hurt after a workout. If they don't
hurt, you didn't work hard enough. If [Student's name] doesn't feel any
pressure, it means he's not learning anything new."
The firm version (for parents who want to negotiate standards):
"Your child is feeling pressure because for the first time they have to
think without outside help. That's not abuse, that's education. If you'd
rather [Student Name] never feel pressure, you're right - I'm not the
right teacher. But if you want [Student Name] to be prepared for a
world where pressure is constant, then the stress now is training, not
punishment."
3.5.4 SYSTEM RESISTANCE (Colleagues, Managers,
Superintendents)
WHAT YOU WILL HEAR:
1. From colleagues: "You're complicating things unnecessarily. Let
them use AI, that's the future anyway."
2. From the principals: "You have a lot of complaints from parents.
Moderate yourself."
3. From the Superintendents: "You have to give X homework per
quarter. The regulations require it."
4. From colleagues: "It's subjective! How do you measure the
'process'?"
HOW TO ANSWER:
Response to objection 1 - "Uselessly complicit"
"I don't complicate, I recalibrate. Yes, AI is the future. But the
future doesn't belong to those who know how to press buttons, but to
those who know how to decide which buttons to press and when not
to press them at all. The difference between an operator (L2) and an
architect (L3) is that the former is replaceable, the latter is not. I teach
them to be L3, not to be L2 or even L1. If you want to train them only
to be operators, I respect that. I want something else for them."
Diplomatic version: "I respect your opinion. But I've seen data
from the first 6 months of introducing AI in education (2023-2024).
Students who used AI intensively have measurable declines in
standardized tests. It's not theory, it's measured. I just apply mental
METEORAIT FOR EDUCATORS 43

safety protocols. If I make a mistake and lose 6 months, we recover.
If you make a mistake and lose 3 years, we don't recover. I prefer to
err on the side of caution."
Response to Objection 2 - "Too many complaints"
Email (formal):
"Dear Management,
I understand that you have received negative feedback from
some parents regarding my assessment methods. I appreciate you
giving me the opportunity to clarify.
Background: We have implemented protocols based on recent
research on the impact of AI on cognitive development (Stan, 2025;
similar studies from the US and EU). These are not personal or
arbitrary methods, but the application of recommended standards to
protect students' cognitive capacity.
Concrete facts: Over the last 8 weeks, I have noticed the following:
- 70% of students show improvement in oral explanation ability
(Micro-checking works)
- Average independent work time (measured in class) increased
from 10 to 20-25 minutes (increased resistance to frustration)
- Complaints come from 15-20% of parents - those whose children
had artificial grades before
Proposal: I am willing to present a session to the School Board about
these methods and why I believe they are necessary. If there is
institutional resistance, I respect the school's decision, but I ask that
the decision be made after presenting the data, not just after emotional
feedback."
Face-to-face conversation: "I know it's uncomfortable to receive
complaints. But the question is: are the complaints legitimate or are
they resistance to increased standards? If a parent complains that the
gym teacher is being too harsh because he makes them run, is the
solution to let them sit on the bench or explain why running is
necessary? I apply the second principle."
Answer to objection 3 - "X homework must be done in X period"
"I understand the requirement. The difference is that the
assessed (graded) assignments are done in class, not at home. The
METEORAIT FOR EDUCATORS 44

homework is preparatory (ungraded or graded with Present/Absent).
This way I respect the bureaucratic requirement, but I evaluate the real
thing. The regulation does not specify where the homework must be
done, only that it must exist."
If the Superintendent insists: "If the regulations require
homework to be done at home and graded, then I ask for confirmation
that the school accepts that grades can reflect the AI's work, not the
students. If they don't accept, then they must allow me to assess under
controlled conditions. I can't do both - respect the form and guarantee
the substance."
Response to Objection 4 - "It's subjective"
"It's less subjective than classic evaluation. When you say an
essay is 'very good', what are you measuring? Your subjective
impression. When I say a process is visible, I have data: “ Version "
History " (objective timestamps), deletions on the sheet (visible), timed
work time. The 'vulnerability' component has subjectivity, right, just like
the 'quality of an argument' in your classic evaluation. But at least now
we're measuring something relevant - learning - not just the final
product that can be generated 100% by AI."
3.5.5 DOCUMENTATION
When applying METEORAIT protocols, build a “protection file”:
1. Initial announcement (written, dated, signed by students /
parents or sent by email): "Starting on date X, the assessment in my
class will be based 50% on the product and 50% on the process. The
process includes: time invested, visible attempts, corrections made,
explanations provided. All students are assessed according to the
same criteria. Using AI at home for preparation is allowed and
encouraged. The final assessment is done in class, independently."
2. Evaluation grids (public, displayed in the classroom and on the
platform): The concrete grids in Chapter 3.4 - so any parent can see
exactly what you are measuring and how.
3. Observation journal (private, for you): You write down: date,
student, what you observed (Micro-check, behavior, progress). If a
complaint comes, you have data, not just impressions.
METEORAIT FOR EDUCATORS 45

4. Comparative results (after 3 months): Measures: how many
students can orally explain what they wrote (at the beginning vs after
3 months), average time to resist frustration (at the beginning vs after
3 months). No scientific study needed, just simple data that shows a
trend.
When the pressure comes, don't defend yourself emotionally. Present
the case: "I announced transparently, I applied consistently, I
measured objectively. Here are the data."
3.6: SELF-SABOTAGE
(When the teacher gives up from fatigue + Fail -Safe Protocol)
3.6.1 THE ELEPHANT IN THE ROOM
Now we're talking about what none of the previous chapters
directly acknowledge: you, the teacher reading this, will give in at
some point. Not "maybe." Not "probably." You will definitely give in.
Not because you're weak. Not because you don't believe in
protocols. But because you're human, you're tired, you have 5 classes,
150 students, 40-50 hours a week, teacher councils, parent meetings,
reports, assessments, bureaucracy, and a personal life waiting for you
at home. And on a Wednesday, at 7 pm, when you still have 40 papers
to correct and tomorrow you have 6 hours, you'll think: "Okay, I'll give
them homework tomorrow and that's it."
That's self-sabotage. It's not dramatic. It's not a big decision.
It's a small, tired concession that you justify with "just this once." The
problem isn't that you're doing it. The problem is that you don't have a
protocol for coming back. And "just this once" becomes "just this
week”, which becomes "just until the holidays”, which becomes "I've
given up completely but I'm pretending I haven't."
This chapter isn't about never giving in. It's about consciously
giving in and coming back without feeling guilty.
3.6.2 WHY EVEN CONVINCED TEACHERS GIVE UP
Reason 1 - Chronic fatigue (not dramatic, but cumulative)
It's not a spectacular burnout. It's a silent exhaustion. Monday:
Micro-check in 3 classes, everything is going well. Tuesday: meeting
METEORAIT FOR EDUCATORS 46

with complaining parents, 2 hours wasted. Wednesday: surprise
Superintendent, stress. Thursday: effort to write assessments in class
when the curriculum requires you to advance with the subject. Friday:
you realize you've corrected half of the papers and you're behind.
On Sunday evening, when you think about the week ahead,
you feel the weight. And then you think, "If I give them homework, I
gain 2 hours. I can correct faster, I can breathe." It's not laziness. It's
the mathematics of survival.
Reason 2 - Isolation (you're the only one in school who does this)
When you apply the METEORAIT protocols and your peers
don't, you become the anomaly. Students complain: "But Mrs. X
leaves us with the phone!" Parents compare: "Why do your grades get
lower?" The principal looks at you strangely: "Why do you have so
many complaints?"
Isolation is corrosive. When all your peers are going one way
and you're the only one going another, you start to doubt yourself.
"Maybe I'm overreacting. Maybe it's just easier to just leave the AI and
be done with it." It's not a lack of conviction. It's social exhaustion.
Reason 3 - Delayed feedback (results are seen in months, not
days)
When you implement protocols, you don't see miracles in the
first month. Students complain, grades drop, parents are unhappy.
Improvement comes after 2-3 months - when students start to think
more clearly, explain better, withstand frustration better. But 2-3
months is an eternity when you're in the middle of chaos.
The human brain is designed to respond to immediate
feedback. "I gave the easy assignment → happy students → I have
peace" = fast reward circuit. "I applied the protocol → unhappy
students → stressed me → results in 3 months" = delayed circuit.
Guess who wins when you're exhausted?
Reason 4 - Lack of community (you have no one to talk to about
it):
You can't talk to your colleagues - they judge you or don't
understand. You can't talk to the principal - he's preoccupied with
complaints. You can't talk to the students - they're the source of
METEORAIT FOR EDUCATORS 47

pressure. You can't talk to the parents - they're the other source of
pressure. You're left alone with your doubts.
When you're alone with a difficult effort, you give in. Not
because the effort is too great, but because the lack of external
validation eats away at you.
3.6.3 SIGNS THAT YOU ARE GIVING UP
Check the behaviors you observe in yourself. 4+ ticks = high risk
of failure:
A - Small compromises:
[ ] "Just this once" I'll let them hand in their homework
[ ] We're skipping Micro-verification for a week, because I don't have
time
[ ] I give them an "easy" homework assignment so they don't complain
anymore
[ ] I accept a vague explanation at Micro-verification because I'm tired
of insisting
B - Justifications:
[ ] "My students are not prepared for such high standards"
[ ] "Maybe I'm exaggerating, maybe it's enough to apply partially"
[ ] "Other teachers don't do this and their students manage"
[ ] "Parents are right, their children really are stressed"
C - Avoidance:
[ ] I avoid checking “ Version History ” because it takes me time
[ ] I avoid giving assessments in class because it delays the subject
[ ] I avoid conflicts with my parents, I accept compromises
[ ] I no longer read materials about FCPT/Cognitive Divergence,
because it demoralizes me
D - Emotional fatigue:
[ ] I think about the coming week on Sunday evening and feel dread.
[ ] I feel resentment towards students when they ask for explanations
[ ] I feel like I'm fighting alone against everyone
[ ] I have fantasies about what it would be like if I gave up and became
like everyone else.
If you checked 4+, don't judge yourself. Just admit it: you give
in. And that's okay. Now comes the return protocol.
METEORAIT FOR EDUCATORS 48

3.6.4 FAIL-SAFE PROTOCOL
(How to come back without feeling guilty)
STEP 1 - Admit the failure (without self-flagellation):
Don't say, "I'm a failure, I gave up, I'm a hypocrite." Say, "I gave
up last week. That's normal. I'm tired. What can I do differently this
week?"
The difference is crucial. Self-flagellation paralyzes you.
Neutral recognition mobilizes you.
STEP 2 - Micro-Rebound (don't try to come back 100% suddenly):
Don't say: "I'm starting everything 100% again on Monday -
Micro-checking daily, in-class assessments, everything." Do: "I'm
starting Micro-checking in one class again on Monday. The rest are
on hiatus for another week."
Gradual recovery is sustainable. Sudden recovery exhausts
you again in 3 days.
STEP 3 - Prioritization (what protocol is essential vs. what is a
bonus):
If you have limited energy, choose what to keep:
Level 1 (Essential - never give up on this):
✓ Micro-verification (1.5 min out of 50, it's minimal)
✓ Ban on clichés (it doesn't cost anything, just announce the list)
Level 2 (Important - keep if you can):
✓ Classroom assessments (at least 1 per month, not 1 per week
if you can't)
✓ “ Version History ” (randomly check 5 students, not all 25)
Level 3 (Bonus - if you still have energy):
✓ Complete protocol (tasks with all 3 methods)
✓ Mandatory self-assessment
✓ Extended Discussions after Micro-Verification
When you give up, don't give up on Level 1. You can let Level 3
“die” temporarily. It's not failure, it's triage.
STEP 4 - Support community (at least virtual):
If you don't have friends at school who do this, search online:
- Discussions with friends, recommended even with those who
have other jobs – objective opinions matter enormously
METEORAIT FOR EDUCATORS 49

- METEORAIT Forum (maybe a Facebook group "Teachers who
apply METEORAIT")
- A single trusted colleague - not from your school, maybe from
another school - who you talk to once a month
You don't need a big community. You just need one person to tell
you, "It's hard, but it's worth it. You're not alone."
STEP 5 - Scheduled Breaks (controlled yielding vs chaotic
yielding):
Instead of giving up when you can't do it anymore, schedule breaks:
Realistic example:
1. Week 1-3: Protocols applied 100%
2. Week 4: " Recovery "week" - protocols applied 50%, you take a
break
3. Week 5-7: 100% Protocols
Or:
1. Quarter 1: Protocols applied 100%
2. Vacation: Zero thinking about protocols
3. Quarter 2: Protocols applied 100%
A scheduled break is not a breakdown. It's maintenance. The car
doesn't go to the MOT/Vehicle Inspection because it's broken, but
because it needs to be done preventively.
3.6.5 WEEKLY SELF-DIAGNOSTICS
(10 minutes Sunday evening)
Take a piece of paper. Write 3 questions. Answer honestly:
1. What did I apply this week from the protocols? Not what I
"should", but what I actually applied.
2. What did I skip/compromise? No self-flagellation. Just the list.
3. Next week: What is that ONE thing I can take back/improve on?
Not 5 things. ONE.
Realistic example:
"1. I applied: Micro-Verification to 2 classes out of 5. I banned
clichés. 2. I skipped: In-class assessments (I gave homework). I did
not verify “Version History". 3. Next week: Resumed classroom
assessments in one class."
After 10 minutes, you close the page. You don't think about it
until next Sunday. Obsession is the enemy of consistency.
METEORAIT FOR EDUCATORS 50

3.6.6 WHEN TO GIVE UP COMPLETELY
Not everything worth starting is worth continuing. There are
situations were giving up is rational, not capitulation:
Situation 1 - The system is actively sabotaging you:
If the principal explicitly tells you: "Stop these methods or I'll
move you to another class / give you a negative evaluation”, and you
have a choice between protocol and a job, choose the job. You're not
a martyr. You're a teacher.
You do what you can within the limits allowed: maybe you can't
apply Flipped Classroom 2.0, but you can apply Micro-checking subtly.
You may not be able to give up homework, but you can ask control
questions.
Active sabotage = you adjust, you don't abandon completely,
but you don't sacrifice your career.
Situation 2 - Personal cost becomes unsustainable:
If after 6 months of consistent application you notice that:
- You have constant anxiety.
- Personal relationships suffer (family, friends)
- Physical health declines (insomnia, recurring illnesses)
- You no longer feel any satisfaction in teaching.
Then stop. No protocol is worth the price of your mental health. It's
not failure. It's self-preservation.
Situation 3 - Zero measurable impact after 1 year:
If after a full year of consistent application, you see no
improvement - neither in the students (they don't think more clearly,
they don't explain better), nor in you (you're no longer sure that you're
assessing realistically), nor in the system (no external validation) -
then perhaps the protocols don't work in your specific context.
Not all methods work everywhere. Maybe your students have
different needs. Maybe your system is too rigid. Don't be stubborn on
principle.
How to quit honestly (without hypocrisy):
If you decide to drop out, tell the students/parents:
METEORAIT FOR EDUCATORS 51

"I tried a new method this year, based on research on AI and
learning. After X months, I noticed that [the results are not what I
expected / the cost is too high / the system does not allow it]. From
now on, I'm going back to more traditional methods. Thanks for your
patience."
Don't pretend that "the methods worked perfectly and now
we're moving on." Recognize the limit. That's maturity, not failure.
3.6.7 MICRO-VICTORY
(How not to forget why you started)
When you're in the middle of fatigue, you forget why you
started. It helps a lot to have a journal of micro-victories. Not dramatic,
not daily. Just when you notice something.
Real examples of what to note:
"Week 3: Andrei, who usually answers 'I don't know' instantly, today
he thought for 15 seconds and then tried an answer. Wrong, but he
tried."
"Month 2: Maria, who always wrote 'perfect' texts (AI), today handed
in an essay with visible erasures and said 'I know it's not perfect, but
it's mine'. I felt something."
"Month 4: Parent who was complaining in October that 'child's grades
are lower' emailed me: 'I've noticed that [Student Name] is explaining
things more clearly at home. Maybe you're right.'"
These are the fuel when you want to give in. You reread them
and remember: it's not in vain.
3.6.8 REALISTIC EXPECTATIONS
(What you can control and what you can't)
You can control:
- Whether to apply Micro-verification or not
- If you check “ Version History ” or not
- If you are assessing in class or at home
- Whether to ban clichés or not
- Your methodological integrity (are you doing what you said you
were doing?)
METEORAIT FOR EDUCATORS 52

- Your reasonable consistency (you apply when you can, you
admit when you can't)
- Accepting your limits (don't pretend you can do more than you
actually can)
You cannot control:
- Whether students like you or hate you
- Whether most students go to L3 or not
- Whether your parents support you or complain
- Whether your colleagues respect you or isolate you
- If the manager promotes you or ignores you
- If the Superintendent understands or penalizes you
- If the education system changes overnight
- If AI disappears or becomes more dangerous
When you feel overwhelmed, read the list above. Let go of what
you can't control. Focus on what you can.
3.6.9 PLAN B
(If you give up, at least don't give up completely)
If you've decided that you can't apply the protocols 100%, there
is a "Plan B" - the minimalist version that's better than nothing:
The minimalist version (15 minutes of extra effort per week):
You keep:
1. Banning 3 clichés (zero effort, just ads)
2. Micro-check 1× per week (not daily), 3 students = 3 minutes
3. At least 1 assessment per month in the classroom (not all)
Temporarily give up:
1. " Version" History "systematic"
2. Mandatory self-assessment
3. Flipped Radical Classroom 2.0
Is the minimalist version perfect? No. Is it better than zero?
Absolutely. Is it sustainable when you're exhausted? Yes.
Golden rule: "The minimalist version applied consistently
beats the maximalist version applied for 2 weeks and then
abandoned."
METEORAIT FOR EDUCATORS 53

PART III – TOOLKIT
This section contains resources that you can use directly,
without major modifications. They are designed to be:
Printable - you can print and display or share
Adaptable - you can change the name, discipline, details
Tested - the structure is designed for real use, not theoretical
You don't have to use them all. Choose what resonates with
your situation. If you only use the FCPT Checklist and the L3
Manifesto, that's more than maybe 90% of teachers do.
RESOURCE 1: FCPT Diagnostic Checklist
(A4 format, printable, for personal use or discussion with colleagues)
DETECTION OF FCPT IN STUDENTS
False Transfer of Cognitive Power - when the student thinks they
know, but they have only witnessed
Instructions: Think about a specific student. Check the behaviors you
frequently observe. 7-8 or more ticks = high risk of FCPT.
SECTION A - Fluency without substance
[...] The student responds quickly and fluently to general questions,
but gets stuck when you ask "Why?" or "How did you get here?"
[...] The written works are stylistically impeccable (advanced
vocabulary, perfect structure, zero mistakes), but the student cannot
explain his own arguments when asked orally.
[...] The sophisticated vocabulary from homework never appears in
spontaneous conversations or class discussions.
[...] When you interrupt him with a side question during an explanation,
he completely loses the thread and can't come back.
SECTION B - Dependence on external support
[...] Can't start a task without "quickly checking something" on the
phone or asking a colleague (euphemism for AI/Google).
[...] Always asks for "an example" before trying it himself, even for
simple tasks that he has theoretically done before.
[...] On tests or assessments without internet access, performance
drops dramatically compared to homework (difference of 3-4 points or
more).
METEORAIT FOR EDUCATORS 54

[...] Shows visible panic or agitation when the phone is not available
or when they have to work completely independently.
SECTION C - Lack of Metacognition
[...] He doesn't know how to say "I'm not sure" or "I don't know" - he
either remains completely silent or tries to feign using generic
formulas.
[...] He cannot realistically estimate how long a task will take him (says
"5 minutes" for anything from a simple exercise to a complex essay).
[...] Doesn't identify errors in his own work when you show it to him
after a few days (doesn't remember the thought process).
[...] When something is wrong in its solution, it doesn't understand
WHY it's wrong, even if you explicitly show it the error.
SECTION D - Asymmetric Performance
[...] Big difference between homework (9-10) and oral quizzes (5-6) or
written tests in class (5-6).
[...] Big difference between written communication (excellent) and oral
communication (confused, hesitant).
[...] The technical/procedural part is perfect (formulas applied
correctly, algorithms followed), but the conceptual part is empty (he
doesn't understand WHY it works).
Interpretation:
0-3 ticks: Student appears to be autonomous. Monitor however -
FCPT may begin subtly.
~ 4-7 ticks: Moderate risk. Student has areas of dependency. Apply
Micro-checking selectively on those areas.
Over 8 ticks: Major risk of FCPT. Student likely delegates thinking
systematically. Urgent intervention: private discussion, intensified
protocols, possible discussion with parents.
What do you do after diagnosis:
You don't blame the student. You don't make the diagnosis
public. You use the information to adapt the assessment and provide
targeted support. FCPT is not moral guilt, it's adaptation to the
environment. The environment must be changed.
METEORAIT FOR EDUCATORS 55

RESOURCE 2: "WHO ARE YOU?" Poster
(A3 format, color, permanently displayed in the classroom)
WHO ARE YOU IN THE AI ERA?
Three strategies. Three destinies.
L1 – THE PASSENGER
Strategy: Full Delegation
What it does: Gives the prompt, gets the answer, sends the
assignment. AI does 80-90% of the work. The student just approves
and “makes cosmetic adjustments.”
How it feels: Efficient, fast, comfortable. "Why sit for 2 hours when I
can finish in 5 minutes?"
Reality in 2 years: Can no longer function without AI. In exams,
interviews, real-life situations without assistance - cognitive paralysis.
Becomes the easiest to replace.
Signs that YOU are here: Can't explain what you wrote. Panic
when the AI is not available. Your homework is "too perfect"
compared to what you can do on the oral exam.
L2 – THE OPERATOR
Strategy: Partial Delegation
What it does: Uses AI to check, structure, generate variants. It
delivers results, but it has lost its intuition. It can no longer detect when
the AI is wrong.
How it feels: Productive, modern, adaptable. "I use technology
intelligently."
Reality in 2-3 years: Competent in standard tasks, stuck in new tasks.
Can operate AI, but cannot critically evaluate output. Vulnerable when
context changes.
Signs that YOU are here: You can explain what you did, but not
why. When AI gives you two options, you don't know which is better
without asking someone else.
L3 – THE ARCHITECT
Strategy: Conscious Symbiosis
What it does: Uses AI to amplify what they already know, not
compensate for what they don't know. AI does the mechanical work,
the student retains conceptual control and the final decision.
METEORAIT FOR EDUCATORS 56

How it feels: Powerful, autonomous, in control. "AI is my tool, not my
prosthesis."
Reality in 2 years: Tens of times more capable than L1. Can operate
with or without AI. Hard to replace. The only sustainable strategy in
the long term.
Signs that YOU are here: You can reconstruct everything you did
without the AI. You can explain why you rejected the solutions the
AI proposed. You feel "cognitive shame" asking the AI trivial things
- it forces you to think first.
QUESTION OF THE DAY:
The last assignment you did - if someone asked you to redo it
completely, without any help, how long would it take you to finish it?
To the same quality?
- If the answer is "at the same time" or "a little slower" → you are
probably L3.
- If the answer is "in double time, but ok" → you are probably L2.
- If the answer is "I couldn't" or "much weaker" → you are probably L1.
It's not judgment. It's reality. And reality can change. Read and
use the tools in METEORAIT FOR TEENS (www.meteorait.eu).
RESOURCE 3: Anti-AI Theme Template
Template 1 - MATH & SCIENCES (Mathematics/Physics)
Topic title: [General topic - ex: "Pythagoras' Theorem"]
LAYER 1 - Informational (AI can solve, 20% of the grade):
Summarize the Pythagorean theorem in 50 words or less. Write the
formula and explain what each element means.
LAYER 2 - Locally applied (AI has no local data, 30% of the
grade):
Find a situation in your home, schoolyard, or neighborhood where the
Pythagorean theorem would help you calculate something you can't
measure directly. Describe the situation specifically (what object, what
distance, what angle) and explain the calculation.
LAYER 3 - Personal (AI cannot simulate, 50% of the grade):
Have you tried the calculation in practice? If so, what was the
difference between the calculation and the actual measurement (if you
were able to measure)? Why do you think the difference occurred?
METEORAIT FOR EDUCATORS 57

If you haven't tried, honestly explain why not (what stopped you, what
you were too lazy to do, why you didn't have time). It's not wrong to
not have tried - I just want to understand your thought process.
BONUS (optional, +1 point):
On a separate sheet of paper, write down 2 questions you asked
yourself while doing your homework, but didn't know how to answer or
couldn't find the answer.
Forbidden clichés (each occurrence = -1 point): "In conclusion",
"Thus", "It is important to mention", "On the one hand... on the other
hand"
Template 2 - HUMANITIES (Native Language / Literature)
Topic title: [General topic - ex: "Characterization of character X from
novel Y"]
LAYER 1 - Informational (AI can solve, 20% of the grade):
Describe in 80-100 words the main traits of character X. Use at least
2 quotes from the text as evidence.
LAYER 2 - Local comparison (AI does not know your context,
30% of the grade):
Compare character X to a person you see often in real life (not a
celebrity - it could be a family member, a neighbor, a store clerk, the
bus driver you take every day). Describe a specific situation in which
character X would react similarly or differently to how the person in
your life reacts.
Example: "Character X, when confronted with injustice, reacts
violently. Mrs. Maria from the store, when a customer is rude, reacts
exactly the opposite - she becomes even more polite, but cold. I think
the difference comes from..."
LAYER 3 - Personal (AI cannot simulate genuine emotion, 50% of
the grade):
How would you react in character X's situation (choose a specific
scene from the novel)? Not what you "should" do or what is "morally
correct”, but what you honestly believe you would do in reality. If you
are not sure, explain why you cannot decide.
BONUS (optional, +1 point):
What argument against your interpretation (about the character or how
you would react) do you find the strongest? I mean, what could make
you change your mind?
METEORAIT FOR EDUCATORS 58

Forbidden clichés (each occurrence = -1 point): "Throughout the
novel", "In contemporary society", "The character is", "In conclusion",
etc.
Template 3 - HUMANITIES (History/Social Sciences)
Topic title: [General topic - ex: "Causes of the Revolution of 1848"]
LAYER 1 - Informational (AI can solve, 20% of the grade):
Summarize in 100 words the 3 most important causes of the
Revolution of 1848. Use at least 2 sources (textbook + one more
source of your choice).
LAYER 2 - Interpretive (AI generates generic, 30% of the grade):
Of the 3 causes you mentioned, which do you think, if they had been
missing, the revolution would not have happened at all? Or do you
think they were all necessary simultaneously? Argue.
Identify a cause that the teacher did NOT mention in class or that the
textbook treats superficially (maximum one paragraph). Explain why
you think it deserves more attention.
LAYER 3 - Personal (AI cannot simulate real biography, 50% of
the grade):
Interview a family member (parent, grandparent, uncle) about a
historical event that they experienced firsthand or about which they
have a strong opinion. Describe: who the person is, what historical
event, what they said (summarize in 3-5 sentences), what surprised
you about what they said.
Then: if this person had lived in 1848, do you think they would have
supported the revolution or been against it? Why (based on the values
you observed in the interview)?
BONUS (optional, +1 point):
Find a contradiction in your position: an argument or because that
weakens what you argued above. You don't need to resolve the
contradiction - just identify it.
Forbidden clichés (each occurrence = -1 point): "Throughout
history", "Since ancient times", "First of all, secondly", "Of course"
METEORAIT FOR EDUCATORS 59

RESOURCE 4: Rapid Assessment Grids
(Traffic Light System - A4 format, 2 grids per page)
PURPOSE: Rapid triage. Do not calculate points if not necessary.
GOLDEN RULE: Start with RED. If you check RED, you stop. Don't
read the rest. The paper is rejected or requires immediate Micro-Oral
Check. Don't waste your time correcting an AI-generated paper.
HOW TO USE THESE GRIDS
1. Scanning (5 seconds): Look at the sheet from a distance. Is it "too
clean"? Is it "too structured"? Suspicion of Red.
2. Cliché Test (10 seconds): Quickly read the first and last
sentences. Does it start with "In modern society..." and end with "In
conclusion..."? Straight Red.
3. Scar Test (10 seconds): Look for the cut mistake. Is there no cut
mistake on the entire sheet? Suspect Red/Yellow.
4. The score:
If it is Green, it gives the maximum possible score for the level of
correctness.
If it's Red, just write on the paper: " Come and explain this to me
orally." (Don't grade it yet).
This method reduces proofreading time by 70% by eliminating time
wasted reading AI-generated texts.
METEORAIT FOR EDUCATORS 60

1. QUICK GRID - MATH & SCIENCES
(Mathematics / Physics / Info)

Student Name: _________________________ Date: ___________

| TRAFFIC  | VISUAL INDICATORS (check      | ACTION /  |     |
| -------- | ----------------------------- | --------- | --- |
| LIGHTS   | what you see in the first 10  | SCORE     |     |
seconds)
|     | - Correct final result, but zero  | STOP.  |     |
| --- | --------------------------------- | ------ | --- |

|     | intermediate steps.  | Don't correct.  |     |
| --- | -------------------- | --------------- | --- |
RED
| (AI /    | - "Alien" method (uses          | Score: 1-4 (or   |     |
| -------- | ------------------------------- | ---------------- | --- |
| Copied)  | formulas not yet taught).       | invite to Micro- |     |
|          | - Suspiciously clean sheet (no  | verification).   |     |
erasures, robotic writing).
|              | - Correct standard solution,       | CHECK.           |     |
| ------------ | ---------------------------------- | ---------------- | --- |
| YELLOW       | logical steps.                     | Read carefully.  |     |
| (Uncertain)  | - Written evenly, without visible  | Score: 5-8       |     |
|              | hesitations.                       | (depending on    |     |
|              | - It seems "too bookish".          | correctness).    |     |
|              | - Erasures, cuts, corrections      | BONUS.           |     |

|     | (traces of the struggle with the  | Reward effort.  |     |
| --- | --------------------------------- | --------------- | --- |
GREEN
|     | problem).  | Score: 9-10  |     |
| --- | ---------- | ------------ | --- |
(Human
|     | - Alternative methods tried  | (even with small  |     |
| --- | ---------------------------- | ----------------- | --- |
thinking)
|     | (even if wrong).            | calculation  |     |
| --- | --------------------------- | ------------ | --- |
|     | - Explanations in your own  | errors).     |     |
words in the margin.

| METEORAIT FOR EDUCATORS  |     |     | 61  |
| ------------------------ | --- | --- | --- |

2.  QUICK  GRID  -  HUMANITIES  (Native  Language  /  Foreign
languages / History)

Student Name: _________________________ Date: ___________

| TRAFFIC  | VISUAL INDICATORS (check what           | ACTION /  |     |
| -------- | --------------------------------------- | --------- | --- |
| LIGHTS   | you see in the first 10 seconds)        | SCORE     |     |
|          | - AI clichés: "In conclusion", "On the  | STOP.     |     |

|     | one hand", etc.  | Don't read  |     |
| --- | ---------------- | ----------- | --- |
RED
|     | - Perfect grammar, but zero  | any further.  |     |
| --- | ---------------------------- | ------------- | --- |
("AI
| Voice")     | substance (empty sentences).           | Score: 1-4      |     |
| ----------- | -------------------------------------- | --------------- | --- |
|             | - Suspicious length (exactly 500       | (Mandatory      |     |
|             | words, 5 fixed paragraph structure).   | retake).        |     |
|             | - Correct, but boring.                 | CHECK.          |     |
| YELLOW      | - Good information, but no personal    | Look for an     |     |
| (Operator)  | opinion.                               | original idea.  |     |
|             | - Standard, safe structure.            | Score: 5-8.     |     |
|             | - Local/personal examples (The         | BONUS.          |     |
| GREEN       | lady at the store, the grandfather).   | Reward your     |     |
|             | - Vulnerability ("I'm not sure, but I  | "own voice."    |     |
(Own
|     | think...").  | Score: 9-10.  |     |
| --- | ------------ | ------------- | --- |
voice)
- Imperfect style (long/short
sentences, human rhythm, passion).

| METEORAIT FOR EDUCATORS  |     |     | 62  |
| ------------------------ | --- | --- | --- |

RESOURCE 5: Email Templates
EMAIL 1 - REPLY TO PARENT (Smaller Note)
Subject: Re: Note on [topic name]
Dear [Parent Name],
Thank you for your email. I understand your concern regarding
[student name]'s grade on a recent assignment.
Let me clarify what has changed in my evaluation method since
[month]:
Before: Students submit homework. I correct and grade the final
product.
Problem: I cannot verify with certainty how much of the assignment
was worked on independently and how much with AI assistance.
Now: I evaluate 50% the final product (quality of the answer) and 50%
the visible process (how the student got there - time invested,
attempts, corrections, explanations).
What does this mean for [student name]:
The assignment had a very good final product (Tg grade: [X]/50), but
the process was almost invisible - "clean" text, no errors, no visible
alternative attempts, very short work time according to "Version
History".
The final grade reflects both components: [(Tg + Tb) / 10].
Why I do this:
According to recent research on the impact of AI on learning, constant
delegation of thinking leads to progressive cognitive atrophy. I don't
want [student's name] to have artificial grades now and discover in 3-
5 years, on a major exam or in a professional context, that the
foundation is fragile.
What we can do together:
I would appreciate discussing how [name] can improve his/her work
process - not for higher grades, but for real learning. I am available
[days/hours] for a meeting.
Respectfully,
[Your name]
[Discipline]
METEORAIT FOR EDUCATORS 63

EMAIL 2 - INITIAL ANNOUNCEMENT TO PARENTS (beginning of
the school year)
Subject: New evaluation method - [Subject] - Class [X]
Dear Parents,
Starting on [date], I will implement a new assessment method in class
[X], in [subject].
Context:
AI tools (ChatGPT, Gemini, Claude, etc.) are now available for free to
all students. They can generate assignments, solve problems, and
write essays in seconds. The reality is that I can't stop students from
using them at home, and it wouldn't be productive to try.
What is changing:
At home, students can use AI freely for preparation (to understand
concepts, practice, check).
At school, the assessment is done independently, without access to
AI/internet/phone.
The final grade will reflect 50% the product (quality of the answer) and
50% the process (how the student thought).
Why I do this:
I want to assess what your child can do independently, not what they
can do with assistance. The difference is crucial for long-term
readiness.
What can you do:
Encourage [student name] to use AI to learn (ask for explanations,
check logic), not to avoid thinking (copy answers).
If you notice that [name] spends very little time on homework (under
20-30 minutes), there may be a problem.
Questions/feedback:
If you have any questions or concerns, please contact me. The
method is new and I appreciate constructive feedback.
Respectfully,
[Your name]
METEORAIT FOR EDUCATORS 64

EMAIL 3 - TO THE DIRECTOR
Subject: Report on evaluation methods - [Discipline]
Dear Management,
I understand that you have received feedback from some parents
regarding my assessment methods in class [X]. I appreciate you giving
me the opportunity to clarify before making any decisions.
Context:
We implemented an assessment system in [month] based on recent
research on the impact of AI on learning (Stan, 2025, available on
SSRN). These are not personal or unfounded experimental methods,
but the application of protocols recommended by specialists in
education and cognitive psychology.
What I did specifically:
- Evaluation 50% on product (quality of response) + 50% on process
(time invested, visible attempts, explanations provided)
- Weekly Micro-Check (3 students, 1 question, 30 seconds) to check
for real understanding
- Homework done in class for final assessment, not at home (where
I can't control the influence of AI)
Observable data after [X weeks]:
- [Y]% of students show improvement in oral explanation ability
- Average independent work time increased from [X] to [Y] minutes
- Complaints come from approximately [Z]% of parents - mostly
those whose children had artificial grades before
Proposal:
I am willing to present a session on these methods to the Faculty
Council (30-45 min). If there is institutional resistance after a full
presentation of the data and methodology, I respect the school's
decision.
I only ask that the decision be made after objective evaluation of the
method, not just based on emotional feedback from parents who are
uncomfortable with the increased standards.
I remain available for any clarifications.
Respectfully,
[Your name]
METEORAIT FOR EDUCATORS 65

EMAIL 4 - TO COLLEAGUES
Subject: Experimenting with new evaluation methods - anyone
interested?
Hello colleagues,
I've started testing some new assessment methods in my
classes, inspired by recent research on AI and learning. I don't claim
to have found the perfect solution, but the early results look promising.
What I do differently:
- I also partially evaluate the "process" (how the student thinks), not
just the final product
- I use "Micro-Checking" - 3 random students, 1 short question, to
see if I understand what they wrote
- Assignments assessed in class, not at home (to eliminate
uncertainty about the influence of AI)
Why I'm sharing this: If anyone is interested in experimenting
together or discussing what works/what doesn't, I'd appreciate it. I'm
not trying to "convert" anyone - I respect that everyone has their own
methods. But if any of my colleagues feel the same frustration with
"too perfect" homework and want to try something different, let's help
each other out.
Resources: I have some homework templates and assessment grids
that I made. If anyone wants them, I'm happy to send them (no
obligation to use them).
Coffee/chat when you have time?
[Your name]
METEORAIT FOR EDUCATORS 66

PART IV - EPILOGUE
WHERE WE ARE
We started from a harsh premise: classic assessment is
"history." Not because teachers were wrong, but because reality has
changed faster than our measurement systems. When a student can
generate a perfect assignment in 2 minutes and there is no reliable
way to detect it, the construct of "homework" collapses.
I have tried to build into this guide not a perfect solution, but a
set of realistic protocols for a real, tired, isolated teacher who wants to
do the right thing but doesn't have infinite energy. It is not a utopian
manifesto about how education should be. It is a guerrilla manual for
how to survive and stay honest in a system that no longer works by
the old rules.
We combined three axes: the student's age (10-14 and 14-18
years old), the type of discipline (Sciences / Humanities), and the level
of effort you can sustain (full protocol & minimalist version).
WHAT THIS GUIDE CANNOT DO
This guide does not save the education system. It does not
transform all schools overnight. It does not guarantee that all your
students will become L3. It does not solve your problems with the
principal who does not understand, with the inspector who demands
"efficiency", with the parents who only want " high grades".
This guide does not give you superpowers. It does not take
away your fatigue. It does not eliminate your 40-50 hours a week of
real work. It does not make you immune to fatigue and exhaustion. It
does not guarantee that you will succeed or that your effort will be
recognized.
This guide cannot tell you „If you do exactly this, you are
guaranteed success". Because there is no such thing. Education is
NOT a first-degree equation (you put in x, you take out y), but a
complex system.
The guide is not even complete. It lacks protocols for special
situations (students with ADHD, anxiety, neuro-divergence). It lacks
extensive empirical data (the theories behind the guide are new and
long-term validation does not yet exist). It lacks the community of
practice that should exist around these ideas, but that we want to build
together.
METEORAIT FOR EDUCATORS 67

WHAT THIS GUIDE CAN DO
It can give you clarity. When a student is turning in a perfect
assignment and you feel like "something is wrong" but you don't have
the words for it - now you have: FCPT. When you notice that some
students are growing exponentially and others are stagnating but you
don't understand the mechanism - now you have: L1/L2/L3 Cognitive
Divergence. Clarity doesn't solve the problem, but it helps you not to
go crazy.
It can give you concrete tools. You don't have to reinvent the
wheel. You have the FCPT Checklist. You have the Micro-verification.
You have Anti-AI homework templates. You have the Tb+Tg
Evaluation Grids. You have email templates for parents / principals /
colleagues. They are not perfect, but they are logically tested and can
be quickly adapted to your situation.
It can give you permission to let go of what no longer works.
Does the system force you to give X number of assignments per term?
Okay, give them, but don't grade them there. Do your parents demand
high grades? Okay, explain to them why the current grade is real and
the old one was fake, but don't give in to the pressure. Do your peers
judge you? Let them, you have data / argument, they just have inertia.
It can validate your intuition. If you already felt that "something
is wrong" with perfect homework, that "students know too little about
what they are teaching”, that "the system rewards bluffing " - you were
not paranoid. You were an observer. Now you have a theoretical
framework that confirms your observation and gives you an answer.
It can protect you institutionally. When a parent complains or
pressure comes from the principal, don't defend yourself with "I
think..." but with "according to the literature on FCPT, the protocol I'm
applying is recommended for..." It's not arrogance, it's armor. The
difference between being seen as "the teacher who complicates" and
"the teacher who applies research-based standards" is enormous.
It can give you a plan for coming back when you give in.
Because you probably will give in at some point. And when you give
in, it won't be the end. It will be a break. And after the break, you have
a fail -safe protocol: Micro-Return, Prioritization Levels 1-2-3,
Supportive Community, Scheduled Breaks. You're not a failure if you
give in. You're a failure if you give in and lie to yourself that you didn't
give in.
METEORAIT FOR EDUCATORS 68

THE LAST WORD
One of the main conclusions of “METEORAIT” is that “
Adulthood is a Process, not a Person.” And the Process is born
from Questions (“ Rogo, ergo emergo ” – I ask, therefore I become).
METEORAIT FOR EDUCATORS is not against AI. It is against
relying on AI without a solid cognitive foundation. You can’t be a good
“ Prompt Engineer ” if you don’t understand the material you’re asking
the AI to process. You can’t detect when AI is wrong if you don’t know
the correct answer yourself. Amplification comes AFTER competence,
not instead of it.
Children are "set" to " here and now " mode. The prefrontal
is immature - the brain's control center, responsible for planning,
impulse control and long-term vision, is fully developed only around
the age of 25. The limbic system dominates - in children, the part of
the brain responsible for emotions and instant gratification is much
more active than the rational part. Stanford Marshmallow Test - the
famous Stanford studies have shown how children have real
difficulties in delaying reward, and those who manage to wait tend to
be more successful in adulthood (but the skill is trained over time). The
concept of time - for a 5-year-old child, "in a week" is an abstract
eternity, they live in a continuous present, which is why the reward
(including the dopamine "for completing the homework") of 5 minutes
later is more valuable than a huge one tomorrow (a grade, for
example) or in the adult future. In short: it's not bad will, it's just
"hardware" that doesn't yet support strategic planning " software".
So perhaps the most important thing you can do is be
consistent in your actions over time. If you're thinking, "It's too much,
I can't do it”, you're probably right. It's probably a lot. And maybe you
can't do everything all the time. But maybe you can do something
consistently.
If you just do the Micro-Check once a week (2-3 minutes out of
50), without anything else from this guide, you have done more than
maybe 90% of teachers in the current system. If you ban just 3 AI
clichés, without changing anything else, you have raised the bar for
those 25-30 students. If you only do one assessment per month in the
classroom, not at home, you have a sample of reality in the middle of
the “ ocean of illusions”.
METEORAIT FOR EDUCATORS 69

You don't have to be a hero. You don't have to save the
system. You just have to be honest in your class, with your students,
with the tools you have. And if out of the 25-30 students you have in
your class, even 5-10 end up being L3 instead of L1, you've changed
something.
If you change the destiny of just one child by what you do
or say, it's like you've lived another life. You probably won't know
who you changed. Or maybe you'll get confirmation in 10 years, or
maybe never. The absence of feedback is not the absence of impact.
You are in the classroom. You are the daily witness to the
difference between learning and simulating learning. The protocols
in this guide are just tools. Use them when you can. Let them go when
you can't. And come back to them when you're ready. Don't judge
yourself. Don't compare yourself to an ideal version of a teacher that
doesn't exist. Compare yourself to yourself 6 months ago. If you have
even one clearer conviction, one more tool, one less failure - you've
made progress.
Ultimately, education is not about AI or anti-AI. It's about the
ability to think critically, to resist frustration, to learn from mistakes,
to remain curious when the answer is a click away. It's about
choosing friction when comfort is easier. It's about being brave
enough to admit "I don't know" and persistent enough to search until
you know.
This book was also written with the help of AIs. More AI
models, more interactions, using information from more sources (for
additional context/understanding). I am a simple engineer, and even
though I graduated from the Teacher Training Program postgraduate
course 20 years ago, what I wrote in this book is about today's reality,
about the exponential development curve of AI and the impact on
present and future human society. AI arrived in our world, each of
us, exactly like a meteorite. It did not ask or ask for permission, and
it cannot be deflected - we can only try not to become fossils (like
dinosaurs) and to adapt (like mammals).
For you, as an educator, I think the best meaning of Rogo, ergo
emergo. is: (When) the teacher asks, the student becomes.
Thank you for reading this far, to the end. With great respect,
Adi Stan.
METEORAIT FOR EDUCATORS 70