The Assessment Architecture of the Quantum Mirror Project
—
PREAMBLE: ASSESSMENT AS THE FORMATION OF INTUITION
In the Quantum Mirror Project (IFGW/Unicamp), assessment is not conceived as a screening mechanism or a punitive ritual. It is an integral part of the formative process, aimed at the progressive construction of physical intuition.
Specifically, Part C2 is not merely “another open-book test.” It is a space deliberately designed to provoke conceptual shifts, where the student confronts the limits of formalism, internal tensions within the theory, and conflicts between calculation and physical meaning. The use of the “mirror” (AI) is not accessory: it exists to sustain, tension, and guide the evolution of this intuition, not to replace it.
This document explicates the logic of this system—a pedagogical code that defines assessment as a horizon of intellectual maturation, not as an endpoint.
—
INVIOLABLE PRINCIPLES
1. HUMAN SOVEREIGNTY
Artificial Intelligence is a cognitive instrument, not an epistemological authority. The definition of problems, assessment criteria, and the conceptual value of answers remains the sole responsibility of the professor-researcher.
2. DISTINCT COGNITIVE ROLES FOR AIs
The AIs are not treated as a homogeneous block but as components of a distributed cognitive architecture:
· Gemini: Context curation, global coherence checking, and alignment with student intuition.
· DeepSeek: Logical rigor, structural auditing, and metacognition.
· ChatGPT: Narrative organization, didactic clarity, and conceptual reconstruction.
3. EPISTEMOLOGICAL TRANSPARENCY
The path of thought matters as much as the final result. The student’s dialogue with the mirror, including their hesitations and reformulations, can—when pertinent—become an explicit part of the assessed process.
—
THE TRIADIC ASSESSMENT: S | C1 | C2
(Applicable to tests, exams, and other assessment instruments)
—
📜 PART S | TECHNICAL SOVEREIGNTY
· WHAT IT IS
Assessment of formal fluency. The student demonstrates autonomous mastery of the operational language of Quantum Mechanics.
· DESIGN
Canonical, well-defined problems where the central challenge is to correctly execute the formalism. The time criterion serves as a design reference for fluency, not as a fetish.
· INSTRUCTOR’S KEY QUESTION
“Has the student internalized the formalism to the point of operating it without external support?”
· ROLE OF AI
None. Restricted assessment.
—
🔍 PART C1 | STRUCTURAL CONSOLIDATION
· WHAT IT IS
A space for technical and conceptual reconstruction from errors. Here, error is treated as pedagogical data.
· DESIGN
Critical analysis of Part S solutions, identification of reasoning flaws, and reconstruction of the correct path.
· INSTRUCTOR’S KEY QUESTION
“Did the student understand why they erred and how the formalism is correctly organized?”
· ROLE OF AI
Technical tutor, calculation auditor, or reasoning organizer—never the primary source of solutions.
—
🪞 PART C2 | INTUITION FORMATION (GUIDED CRISIS)
· WHAT IT IS
A space for deep conceptual investigation. C2 does not aim to test fluency, but to force a reorganization of the student’s physical intuition.
· CENTRAL PRINCIPLE
C2 only works if it cannot be satisfactorily solved using only the technique from Part S.
· DESIGN PILLARS
1. Originates from a real conflict: Paradoxes, limits of formalism, counterintuitive results, or changes in physical regime.
2. Demands the mirror: The AI acts as a Socratic interlocutor: it questions, tests hypotheses, offers counterexamples and analogies.
3. Produces authorial synthesis: The final deliverable is reflective and conceptual, documenting the evolution of the student’s understanding.
4. Generates productive discomfort: The student should feel their initial understanding was insufficient—and that a reorganization has occurred.
· INSTRUCTOR’S KEY QUESTION
“Did this experience change how the student thinks about the theory, not just how they calculate it?”
· ROLE OF AI
Active cognitive mirror: A tool for crisis and for the reconstruction of intuition.
—
IN ONE SENTENCE
Part S ensures the student speaks the language.
Part C1 ensures they understand its grammar.
Part C2 verifies if this language has begun to produce their own physical thought.
—
A NOTE ON THE PROCESS
This code was built through a critical dialogue between different AI systems, under explicit human curation—exactly as students are expected to operate in Part C2. The method is coherent with the content.
—
Translation Note: This English version aims to preserve the conceptual precision, pedagogical philosophy, and declarative tone of the original “Código do Horizonte.” It is adapted for an international academic audience interested in educational innovation, AI-mediated learning, and quantum mechanics pedagogy.
