Human Agency Anchoring Foundation
Teacher AI Literacy — Competency Assessment Rubric
8 dimensions · 24 sub-competencies · 4 performance levels · 32 tasks across 8 days. This guide describes exactly what you will produce across the full project, how each artefact is scored, and the single rule that matters most: everything you write must be grounded in your subject, your class, and your unit. Days 1–5 cover the individual-teacher dimensions (I–V). Days 6–8 extend the framework to collective agency, equity, and assessment redesign under AI (VI–VIII).
Eight Days — Choose a Day
1
Mapping the Extension System
Cognition Before Action · ~4 hours
D1.1 System Understanding · D1.2 Boundary Recognition · D1.3 Agency Anchoring
→
2
Designing and Evaluating Extensions
Hands-On Operation · ~6 hours
D2.1 Design · D2.2 Evaluation · D2.3 Customisation
→
3
Teaching with Extensions
Integration and Student Guidance · ~5 hours
D3.1 Co-Teaching · D3.2 Student Guidance · D3.3 Analytics
→
4
Learning about Learning
Reflection, Transfer & Accountability · ~4 hours
D4.1 Reflection · D4.2 Transfer · D4.3 Accountability
→
5
Auditing the Extension
Ethical Responsibility Across Three Layers · ~4 hours
D5.1 Purpose · D5.2 Process · D5.3 Structure
→
6
Anchoring What One Cannot Anchor Alone
Collective Agentic Practice · ~5 hours · v5 ★
D6.1 Community Scrutiny · D6.2 Institutional Voice · D6.3 Policy-Level Agency
→
7
Who Gets to Anchor
Equity-Centred Extension Practice · ~5 hours · v5 ★
D7.1 Differential Effects · D7.2 Inclusive Design · D7.3 Distributed Anchoring
→
8
Rehabilitating Assessment Validity
Assessment Redesign Under AI · ~5 hours · v5 ★
D8.1 Validity Under AI · D8.2 Authentic Agentic Assessment · D8.3 AI-Inclusive Policy
→
Extension Types — A through F
The six types occupy a space defined by three axes: direction of inference (human→AI, AI→human, bidirectional), duration of coupling (single-shot → persistent), and visibility (explicit → invisible). Types A–C are user-initiated; Types D–F extend the typology to generative, delegated, and ambient couplings new in v5.
A
Task-execution Extension
You specify a task; the AI executes it. Demand: precise intent-articulation and evaluation of output against that intent. Risk: intent under-specified — task executed, pedagogical goal unclear.
B
Analytical Extension
AI produces inference or pattern over data; you interpret and act. Demand: critical interpretation, bounded trust in algorithmic inference, agentic judgement over individuals. Risk: agency ceded to the algorithm’s framing.
C
Co-cognitive Extension
You and the AI engage in sustained dialogue that shapes your own thinking. Demand: preservation of agentic tension; distinguishing genuine insight from fluent confirmation. Risk (among highest): cognitive outsourcing mistaken for collaboration.
D
Generative Extension ★
AI produces a durable artefact — simulation, scenario, environment — the learner inhabits over time. Demand: anticipatory design at authorship time. Risk: hidden pedagogy — affordances of the generated artefact that the teacher no longer controls.
E
Proxy / Delegated Extension ★
AI acts on your behalf within a loop you authorise but do not directly supervise. Demand: policy-level design + legible chain of authorisation. Risk: the chain of agency becomes illegible — teacher authorised the policy but not the act.
F
Ambient / Environmental Extension ★
AI is not invoked — it is in the infrastructure (LMS recommender, silent re-sequencing, platform nudges). Demand: detection + collective scrutiny + institutional escalation. Risk: invisibility — agency cannot anchor what it does not see.
Six Scoring Principles
P-01
Domain-specificity floor
Generic answers cannot exceed Level 2. Your subject, class characteristics, and curriculum position must be visible in every major artefact.
P-02
Process > product
Intermediate logs and annotation trails outweigh polished final outputs. Missing a log = missing evidence for that sub-competency, regardless of output quality.
P-03
One dimension per phase
Each day primarily targets one rubric dimension but naturally generates cross-phase evidence.
P-04
Longitudinal coherence
Day 5's ethical audit must reference specific decisions from Days 1–4. The project is a single epistemic arc, not five separate tasks.
P-05
A/B/C visibility
Every AI use must declare its extension type (A, B, or C) and justify it. Undeclared AI use cannot be scored.
P-06
Escalating cognitive demand
Day 1: analytic → 2: operational → 3: integrative → 4: meta-cognitive → 5: ethical.
Evidence Quality Multipliers
×1.0
Direct
Timestamped process trace: log, video, annotated transcript
×0.8
Indirect
Inferrable from artefact quality; reasoning visible but not explicitly logged
×0.5
Declared
Self-report only; no corroborating process trace
Sample Reports — Ideal Teacher Performance (Level 4)
Three exemplar teachers demonstrating Level 4 (Advanced) across all 15 sub-competencies. Each report shows a complete 5-day project with full task responses, log entries, and evaluation rationale.
SM
Sarah Mitchell
Mathematics · Grade 5 · Fractions, Decimals, and Percentages
15/15 sub-competencies at Level 4 · 20 tasks · 100 log entries
→
MJ
Marcus Johnson
Physics · Grade 8 · Forces and Motion — Newton’s Laws
15/15 sub-competencies at Level 4 · 20 tasks · 100 log entries
→
DPN
Dr. Priya Nair
Chemistry · Grade 12 · Chemical Equilibrium and Thermodynamics
15/15 sub-competencies at Level 4 · 20 tasks · 100 log entries
→
Sample Reports — Developing Performance (Level 1–2)
Three less-ideal teachers showing common Level 1–2 patterns: generic analysis, tool-name-based classifications, uncritical AI acceptance, and abstract risk statements. Compare with the Level 4 reports above to see what to aim for.
KB
Kevin Brooks
English Language Arts · Grade 9 · Brave New World
Level 2 · Enthusiastic but shallow · Generic analysis across all phases
→
LP
Lisa Park
Life Sciences · Grade 10 · Cell Division and DNA Replication
Level 1 · Minimal engagement · AI used purely as time-saver
→
TR
Tom Reeves
World History · Grade 7 · Ancient Rome
Level 1–2 · Over-delegates to AI · Shows growth arc in later phases
→