walk the AI
The competency framework
Published: August 13, 2026 · Version 3
This is what separates the diagnostic from a magazine quiz: an explicit competency framework, anchored in the teaching profession, whose construction method is open to scrutiny. Nothing about a teacher's result is arbitrary or hidden.
What it measures — and what it doesn't
The diagnostic measures a teacher's or trainer's command of AI in their own work — not general knowledge about artificial intelligence. Someone may not know how a neural network works and still run an exemplary use of AI in the classroom; it is that use the framework assesses. It rests on five domains, each scored from 0 to 100.
| Letter | Domain | What it assesses |
|---|---|---|
| A | Mindset & attitude | Critical curiosity, evaluating tools against real tasks, keeping professional judgement |
| B | Ethics & responsible use | The framework, pupil-data protection, verification, transparency, safeguarding, age & equity of access |
| C | Fundamentals & tools | How the tools work, writing instructions, knowing tools, checking outputs |
| D | Augmented teaching | Lesson design, assessment, differentiation, inclusion — AI assists, the teacher decides |
| E | Professional growth | Training, sharing, reliable sources, a strategy for progress |
Alignment with the UNESCO framework
The five domains map directly onto the five dimensions of UNESCO's AI Competency Framework for Teachers (2024), and our three levels are UNESCO's three progression levels, word for word: Acquire → Deepen → Create. This is an independent self-assessment built on that public structure — not the official UNESCO instrument.
| This framework | UNESCO AI Competency Framework for Teachers |
|---|---|
| A — Mindset & attitude | Human-centred mindset |
| B — Ethics & responsible use | Ethics of AI |
| C — Fundamentals & tools | AI foundations & applications |
| D — Augmented teaching | AI pedagogy |
| E — Professional growth | AI for professional learning |
Reference: UNESCO, AI Competency Framework for Teachers (2024). An independent alignment, not an official UNESCO scoring instrument.
The competencies, domain by domain
A — Mindset & attitude
- Approaching AI with curiosity and critical thinking, rather than avoidance or naive adoption.
- Evaluating a new tool against concrete tasks, without waiting for it to be imposed.
- Being able to explain simply what AI can and cannot do in the profession.
- Seeing where AI saves time without harming teaching quality.
- Holding a sound line on "replacement": AI automates tasks, not the teaching relationship.
- Keeping ownership of the decision when AI makes a recommendation about a pupil.
B — Ethics & responsible use
- Knowing the framework (school rules, legal context) governing use with pupils.
- Protecting personal data: minimisation, anonymisation, no identifying data.
- Spotting and discussing errors and biases ("hallucinations").
- Addressing honest, critical use with pupils (integrity, plagiarism).
- Verifying at source the references an AI supplies with confidence.
- Owning and explaining one's own uses to pupils.
- Keeping AI out of sensitive decisions (safeguarding).
- Framing direct pupil use: approval, the minimum age in the terms, an alternative for those without an account.
- Accounting for equity of access: not every pupil has AI at home.
C — Fundamentals & tools
- Understanding, without technical detail, that an assistant predicts the most plausible text — with no built-in notion of true or false.
- Writing a useful instruction ("prompt"): role, context, audience, format, example.
- Knowing and using tools suited to education.
- Systematically checking and correcting an output before using it.
- Knowing where vigilance is highest: calculations, precise figures, exact quotations.
- Respecting rights, licences and disclosure of AI origin before distribution.
D — Augmented teaching
- Designing lessons, materials and differentiated activities with AI.
- Designing assessments, rubrics and feedback.
- Differentiating while keeping the pedagogical decision (calibrating and validating variants).
- Building pupils' critical thinking on AI-produced work.
- Evolving assessment: also assessing the process (drafts, oral, declared use).
- Using AI for accessibility and inclusion (dyslexic, non-native pupils) without dropping the learning demand.
E — Professional growth
- Training regularly on AI uses in one's profession.
- Sharing practices with colleagues or a community.
- Knowing where to find reliable resources.
- Having a clear sense of what to learn next.
- Adopting a durable strategy: one precise goal applied to one's tasks, tested, with peer feedback.
- Being able to guide a beginner colleague, starting from their real tasks.
The eight archetypes — three axes
The individual profile is read on three axes, each split at 50/100. Their combination gives one of eight archetypes (presented as falcons).
| Axis | Computed on | Two poles |
|---|---|---|
| Technical ease | (Fundamentals + Growth) / 2 | Intuitive (I) ↔ Equipped (O) |
| Pedagogical integration | (Teaching + Mindset) / 2 | Cautious (P) ↔ Driver (M) |
| Ethical stance | Ethics | Trailblazer (D) ↔ Guardian (G) |
| Archetype | Combination | In one line |
|---|---|---|
| The Enlightened Pioneer | O·M·G | The complete profile: mastery, integration, ethical awareness |
| The Trailblazer | O·M·D | Bold and at ease; the framework is the growth area |
| The Vigilant Strategist | O·P·G | Solid and impeccable; only the pedagogical leap remains |
| The Tech Explorer | O·P·D | Agile with tools; anchor it in pedagogy and a framework |
| The Devoted Educator | I·M·G | Excellent in the classroom; a few fundamentals to consolidate |
| The Curious Enthusiast | I·M·D | The drive is there; the task is to structure it |
| The Thoughtful Sentinel | I·P·G | Rare discernment; it only needs the right tools |
| The Fledgling Scout | I·P·D | Everything lies ahead; no bad habits |
How the score is computed
- Each answer is worth 0 to 4 points.
- Domain score = points earned ÷ maximum possible × 100.
- Overall score = the average of the five domains (each weighs equally).
- Three levels: Acquire (0–40), Deepen (41–70), Create (71–100).
Points are never shown to the respondent: they see their score per domain, their level and their archetype.
How it was built — and what it is not
In the interest of honesty: this is not a peer-reviewed, externally validated psychometric instrument, and we do not present it as one. What it is:
- 44 questions (43 scored + one open feedback question), at strict parity between French and English.
- Two complementary formats: behavioural items ("Since the start of term, what have you actually done…"), which measure a fact rather than a self-opinion; and scored scenarios, with "expert-sounding" distractors to separate the strongest.
- Criterion-referenced scoring, not self-rating. We don't ask "can you do this?" (everyone over-rates) but "faced with this situation, what do you do?" — which curbs social-desirability bias.
- Discrimination at the top: near-miss options worth 3 points separate the good from the excellent.
- A stable scale: revisions preserve comparability over time; the thresholds (50 for the axes, 40 and 70 for the levels) stay fixed.
- Anchored in the profession: every item is written in a teacher's daily reality, never in the technological abstract.
- No email wall: the result shows without an account; an email is asked only to save or to unlock the report.
The five domains cover the ground recognised references treat as core — ethics and data protection, pedagogy, fundamentals, professional growth. We would welcome scrutiny of the framework, and are glad to map it against a recognised reference (such as UNESCO's AI Competency Framework for Teachers, or the EU's DigCompEdu) on request. How the scores are kept confidential is set out in Score privacy.