The "Columbo Teacher"

Introduction

In the TV series, Lieutenant Columbo generally knows the culprit from the start. His job is not to discover "whodunit," but to build the proof, track inconsistencies, and demonstrate guilt. This is precisely the mindset we must cultivate in our students when facing AI.

1. Lessons from the Past: When Technology Shifts the Stakes

The Example of Variation Tables

Before graphing calculators, variation tables were the culmination of a long analytical process—an indispensable step for manually plotting a function's curve. The plot itself was part of high school exit exams.

The arrival of technology reversed this logic. The curve became an immediate "given." Plotting disappeared from exams. Yet variation tables didn't vanish; their role transformed: they now serve to synthesize and give meaning to an already-visible object.

What AI Changes Fundamentally

Generative AI pushes this logic to its extreme. It doesn't just provide graphs: it offers complete reasoning chains, structured proofs, argued solutions.

Within Brousseau's Theory of Didactical Situations, the "milieu"—the antagonistic system the student confronts—gains a new element: the machine-generated answer, which must be analyzed, validated, or refuted.

2. From "Sherlock Mode" to "Columbo Mode"

The Traditional Model: The Student-Investigator

Historically, students operated in "Sherlock Holmes" mode. They started from a mystery—the problem—to search alone for the solution. The teacher was the "revealer of enigmas" and the sole source of explanations.

The New Model: The Student-Validator

AI changes the game by providing immediate access to solutions. Faced with an already-generated answer, students must adopt Lieutenant Columbo's stance: build the proof, track inconsistencies, demonstrate.

The pedagogical challenge shifts from raw production to critical analysis and validation of machine-proposed solutions.

Skills of "Columbo Mode"

3. The Teacher's Metamorphosis

From Transmitter to Orchestrator

The teacher doesn't disappear—quite the contrary. They become a "director of learning environments." Learning doesn't end where AI gives an answer. It begins precisely there.

Three New Axes of the Profession

What Remains Irreducibly Human

• Presence - observing a furrowed brow, sensing a classroom
• Relationship - the climate of trust, the reassuring smile
• Intelligence of the moment - changing strategy in a split second

4. The Student in Critical Dialogue with the Machine

AI as "Milieu" in Brousseau's Sense

Far from rendering students passive, AI creates new didactic "milieus":

• AI proposes answers that students must evaluate
• It can be wrong, creating learning opportunities
• It can be queried iteratively

Constructive Conversation

From Reproducer to Judge

Bloom's taxonomy takes on full meaning: lower levels (remember, apply) can be delegated to AI, while higher levels (analyze, evaluate, create) become the core of students' intellectual activity.

5. Epistemological Implications

AI Doesn't Know—It Predicts

This fundamental distinction must be explicitly taught. AI doesn't understand mathematics; it recognizes statistical patterns in language.

The New Status of Error

AI error becomes pedagogical gold. Unlike a textbook (presumed correct), AI can confidently assert falsehoods. This creates unprecedented teaching opportunities.

Pedagogical shift: Error is no longer just student error to correct, but machine error to detect—a more engaging and less threatening dynamic.

7. Assessment in the AI Era

What to Assess?

Not the final product (potentially AI-generated) but:

• The validation process: How did the student verify AI output?
• Critical analysis: What weaknesses did they identify?
• Iterative refinement: How did they improve the machine's response?

New Evaluation Formats

8. Broader Societal Stakes

Beyond Mathematics

While developed for mathematics education, this approach extends to all disciplines. The "Columbo" mindset—critical validator rather than passive receiver—is essential across the curriculum.

Preparing Citizens, Not Just Students

In a world where AI generates content en masse, the ability to critically evaluate machine outputs becomes a fundamental civic competency. We're not just teaching math; we're forming critical citizens.

Conclusion: The Irreplaceable Human Touch

AI doesn't threaten the teaching profession—it redefines it. The teacher's role becomes more essential, not less:

• More curatorial — selecting quality among abundance
• More relational — the human connection AI cannot provide
• More metacognitive — guiding reflection on learning itself
• More critical — modeling skepticism and validation

The "Columbo Teacher" doesn't compete with AI. They teach students to dance with it—skillfully, critically, and autonomously.

Theoretical References

This reflection draws upon 18 frameworks from educational research, including:

• Brousseau's Theory of Didactical Situations
• Rabardel's Instrumental Approach
• Bloom's Taxonomy (revised)
• Wineburg's Historical Thinking
• Paul & Elder's Critical Thinking Framework
• Vygotsky's Zone of Proximal Development
• And others (complete bibliography available in the downloadable PDF)