Lisa Park

HAA Teacher AI Literacy — Performance Report
Subject: Life Sciences
Year Level: Grade 10
Class: 35 students
School: Greenfield Secondary
Unit: Cell Division and DNA Replication
Report Date: 2026-04-18
13
Overall Score
25
D1: Foundations of
33
D2: Extension Operation
25
D3: Extension-Integrated Teaching
25
D4: Extension-Enabled Professional
25
D5: Extension Ethical
25
D6: Collective Agentic
25
D7: Equity-Centred Extension
25
D8: Assessment Redesign

Complete Project Summary

Project Summary: Lisa Park — HAA AI Literacy Project ## Grade 10 Life Sciences: Cell Division and DNA Replication | South Africa, CAPS curriculum

Lisa Park's five-day HAA project illustrates the Level-1 profile with honesty rather than pretension: a teacher who is competent in her subject, uses AI for practical efficiency, but has not yet developed a conceptual framework for what AI does in the teaching-learning relationship. Her work is consistently minimal, generic, and tool-focused, with almost no subject-specific pedagogical reasoning.

Day 1 was the most revealing. Lisa's Extension System Analysis reduced five AI uses to their simplest descriptions: 'making worksheets,' 'making PowerPoint slides,' 'marking assistance,' 'online quiz,' and 'chatbot for questions.' None of these names a pedagogical function — they name production tasks. The A/B/C classifications are assigned but never justified in cognitive terms. The 'If AI were removed?' column is filled with variants of 'I would do it myself' and 'same thing really,' suggesting Lisa sees AI exclusively as a time-saver rather than as something that changes the nature of the activity. Her Boundary Map is the project's thinnest artefact: all five entries are one-sentence observations ('AI can give wrong information,' 'AI can't do practicals,' 'AI doesn't know my students') with generic lesson activities ('All lessons,' 'Lab work,' 'Teaching') and single-sentence justifications. No entry names a specific biology concept, a specific lab activity, a specific student error pattern, or a specific CAPS content standard. The Teacher Irreplaceability Declaration follows the same pattern — all five reasons (reading body language, maintaining discipline, caring about students, adapting on the fly, assessing understanding) are generic teacher capabilities with no connection to biology as a discipline. Lisa's Phase 1 work would read identically if she taught mathematics, geography, or any other subject.

Day 2's Prompt Engineering Log was functional but pedagogically thin. Eight iterations moved from 'make a worksheet on mitosis' to a usable product, but the revision rationale at each stage noted format issues ('too easy,' 'need harder questions,' 'fix diagram issue,' 'add CAPS alignment') rather than anything about what the worksheet needed to achieve in terms of student understanding of cell division. The sixth iteration is notable: Lisa prompted for CAPS alignment but then admitted 'I need to check the CAPS alignment myself' and 'can't fully trust the curriculum alignment' — an honest acknowledgment that the AI doesn't reliably know the South African curriculum, but she did not explore what that means for her practice. The Output Evaluation Reports found a genuine sequencing error (prophase events out of order), an overconfident absolute claim ('always occurs in the S phase'), and a plant/animal cell omission — these are all real biology errors, but the 'why it matters' explanations stopped at 'students could learn the wrong order' and 'students should understand that biology has exceptions' rather than naming the specific conceptual model being disrupted or the downstream learning impact. Her Customised Artefact (Mitosis Worksheet Generator) was a format template with no embedded pedagogy.

Day 3's micro-lesson was structured as present → practice → review, with AI used at each stage in its most conventional role: slides for content delivery, worksheet for practice, memo for marking. The lesson structure was unchanged by AI presence — the technology accelerated production but did not change anything about the learning experience. Lisa's Decision Log entries are terse and observational ('used AI slides for note-giving — standard, nothing special'; 'students were copying notes'). The most significant entry is the exclusion at 25:00: 'Did Q&A without AI — just me and the students talking about what confused them,' followed by the retrospective 'This was the most productive part of the lesson actually.' Lisa noticed this but did not explore what it means — why was the AI-free segment more productive? What does that imply about the AI-mediated segments? The student work evaluations showed one genuine strength: Lisa caught that a rigid AI answer key penalized a student whose answer was actually more nuanced ('Your answer is actually more accurate than the memo!'). But this insight remained isolated rather than informing a broader reflection on AI's limitations in assessing biological understanding.

Day 4 was the project's weakest phase. Lisa accepted all five AI insights about her practice without modification or rejection, and none of her acceptance reasoning went beyond agreement ('True,' 'Yes,' 'Fair'). The AI insights themselves were generic praise ('strong focus on factual accuracy,' 'your approach shows growth in AI literacy,' 'your ethical awareness is developing'), and Lisa did not notice the echo-chamber pattern. Her stakeholder communications used nearly identical content and framing across all three audiences. Her PD goals were entirely tool-focused: 'learn to use AI for creating better visual materials,' 'explore AI tools for lab safety demonstrations.' No goal addressed a specific pedagogical limitation identified in Phases 1-3.

Day 5's ethical audit listed practical concerns (curriculum misalignment, content errors, over-reliance, internet dependency) rather than ethical risks. 'AI tools require internet access which isn't always reliable' and 'Free AI tools might start charging' are infrastructure risks, not ethical analyses. The structural ethics layer produced no analysis of how the AI tools' design assumptions might shape the teaching of biology or affect particular student populations. The Purpose Drift Audit identified efficiency-driven decisions honestly but the revisions were minimal: 'cross-reference with CAPS,' 'at least 50% of materials must be teacher-created,' 'generate and print materials in advance.' No revision changed a tool or platform; none referenced a specific biological concept, student group, or pedagogical principle.

Lisa's project is not the work of a bad teacher — her subject knowledge shows in the specific errors she catches, and her honesty about what works ('the Q&A without AI was the most productive part') suggests genuine reflective capacity. What is missing is a framework for thinking about what AI does in the relationship between her biology expertise and her students' learning. She uses AI as a photocopier with better formatting, and the project did not shift her beyond that.

Evaluation Summary

Lisa Park demonstrates Level 1 performance across most sub-competencies. Work is largely generic — A/B/C classifications reference tool names rather than cognitive functions, boundary maps list general limits without connecting to specific Life Sciences activities, and irreplaceability reasons could apply to any teacher in any subject. The strongest work appears in output evaluation (Phase 2) where Life Sciences-specific errors are identified, and in real-time classroom adjustments (Phase 3) where genuine pedagogical judgment is visible. The weakest areas are self-analysis (Phase 4), where all AI insights are accepted without critical annotation, structural ethics (Phase 5) with risks stated as general principles, and Phases 6–8: Phase 6 (collective practice) stays at parallel individual opinions rather than joint work; Phase 7 (equity) uses generic "diverse learners" framing rather than named groups; Phase 8 (assessment validity) flags the AI-substitution risk but does not redesign the assessment structurally. Evidence is predominantly declared (self-reported) rather than direct (process traces, timestamped logs, video).
Strengths
Honest self-assessment — acknowledged limitations and efficiency-driven decisions openly
Some genuine Life Sciences-specific analysis in output evaluation reports
Real-time classroom adjustments showed authentic pedagogical judgment
Prioritised Growth Areas
Ground every A/B/C classification in the cognitive function being extended, not the tool name — what changes in the teacher's pedagogical thinking, not just the workflow
Connect all boundary analysis to specific Life Sciences concepts, lesson activities, and student populations — generic limits score Level 1
Develop critical stance toward AI self-analysis — identify echo-chamber effects where AI reflects the teacher's own language back as validation
Move from listing risk categories to naming causal mechanisms: which students, in which activity, through what pathway, with what consequence

Rubric Scores — All 15 Sub-Competencies

D1 Foundations of Extended Cognition
D1.1
Extension-System Understanding
A/B/C foundational · Phase 1 · Extension System Analysis + Vendor Annotation
Lisa Park's work for D1.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D1.2
Recognition of Extension Boundaries
A/B/C foundational · Phase 1 · Boundary Map
Lisa Park's work for D1.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D1.3
Anchoring of Teacher Agency
Critical in C-type · Phase 1 · Teacher Irreplaceability Declaration
Lisa Park's work for D1.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D2 Extension Operation Capability
D2.1
Extension Design Capability
Core for A-type · Phase 2 · Prompt Engineering Log (≥8 iterations)
Lisa Park demonstrates basic competency in D2.1 with some reference to Life Sciences content, but analysis remains at the level of tool use rather than pedagogical function. Evidence is primarily declared rather than direct.
Evidence: Declared (×0.5)
Level 2 — Developing
D2.2
Extension Evaluation Capability
Universal, esp. C-type · Phase 2 · Three Output Evaluation Reports
Lisa Park's work for D2.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D2.3
Extension Customisation Capability
Core for B/C-type · Phase 2 · Customised Extension Artefact
Lisa Park's work for D2.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D3 Extension-Integrated Teaching Practice
D3.1
Implementation of Human–AI Collaborative Teaching
A+B+C combined · Phase 3 · Video + Live Decision Log
Lisa Park's work for D3.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D3.2
Guidance and Support for Students' Extension Use
Central C-type scenario · Phase 3 · Guidance Episode (video) + Student Work Evaluations
Lisa Park's work for D3.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D3.3
Application of Learning Analytics
Core B-type capability · Phase 3 · Student Work Sample Evaluations
Lisa Park's work for D3.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D4 Extension-Enabled Professional Development
D4.1
AI-Assisted Professional Reflection
Central C-type scenario · Phase 4 · Self-Analysis with Annotated AI Transcript
Lisa Park's work for D4.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D4.2
Extension-Transfer Capability
Universal meta-capability · Phase 4 · New Tool Evaluation Report (timed)
Lisa Park's work for D4.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D4.3
Extension Accountability Communication
B/C-type application · Phase 4 · Three Stakeholder Communications
Lisa Park's work for D4.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D5 Extension Ethical Responsibility
D5.1
Ethics of Extension Purpose
High-risk in C-type · Phase 5 · Purpose Drift Audit + Revision Log
Lisa Park's work for D5.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D5.2
Ethics of Extension Process
Universal across A/B/C · Phase 5 · Ethical Risk Memo
Lisa Park's work for D5.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D5.3
Ethics of Extension Structure
Deeper B/C-type risks · Phase 5 · Ethical Audit + Revision Log
Lisa Park's work for D5.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D6 Collective Agentic Practice
D6.1
Professional-Community Scrutiny
Primary site for Type F; relevant to E · Phase 6 · Professional-Community Evaluation (joint artefact)
Lisa Park's work for D6.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D6.2
Institutional Voice and Escalation
Core for Types E/F · Phase 6 · Type-F Audit Memo + Institutional Escalation Draft
Lisa Park's work for D6.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D6.3
Policy-Level Agency
Cross-cutting; primary for Type F · Phase 6 · Policy-Level Response
Lisa Park's work for D6.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D7 Equity-Centred Extension Practice
D7.1
Recognition of Differential Extension Effects
General across A–F · Phase 7 · Differential-Effect Map
Lisa Park's work for D7.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D7.2
Inclusive Extension Design
Central for Types A/C/D · Phase 7 · Inclusive Redesign + Training-Data Audit
Lisa Park's work for D7.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D7.3
Distributed Anchoring Capacity
Cross-cutting; primary for Type C · Phase 7 · Anchoring-Capacity Cultivation Plan
Lisa Park's work for D7.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D8 Assessment Redesign Under AI
D8.1
Assessment Validity Under AI
General A–F; acute for A and E · Phase 8 · Assessment Validity Audit
Lisa Park's work for D8.1 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D8.2
Authentic Agentic Assessment
Core for Type C; general A–F · Phase 8 · Authentic Agentic Assessment Design + Revalidation Trial
Lisa Park's work for D8.2 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent
D8.3
AI-Inclusive Assessment Policy
Cross-cutting; primary for A and C · Phase 8 · Co-articulation Record + Shared Policy
Lisa Park's work for D8.3 is generic — it could apply to any subject rather than being grounded in Life Sciences. Evidence is self-reported without process traces or specific examples from the unit.
Evidence: Declared (×0.5)
Level 1 — Nascent

Evidence Log Summary

100 log entries across 6 types

16
LOG-A
Prompt Engineering
11
LOG-B
HAA Classification
16
LOG-C
Live Decision
20
LOG-D
Output Evaluation
22
LOG-E
Ethical Audit
15
LOG-F
Reflection Annotation

Task Work

Click any task to view the full draft in the workspace wizard

Day 1 Mapping the Extension System D1 · Foundations of Extended Cognition
01
Extension System Analysis
500–700 words, no AI access. Classify ≥5 teaching activities from the unit into A/B/C extension types. Each classification must cite the specific learning objective, class level, and what the teacher would do if AI were removed.
Draft · 1,849 ch Submitted No AI access
02
Subject-Specific Boundary Map
Three-column table — technical limits, contextual limits, value-judgment limits. Each row names a specific AI capability, a specific lesson activity, and why the limit matters in this subject at this level.
Draft · 1,042 ch Submitted
03
Teacher Irreplaceability Declaration
≥5 decisions in this unit that cannot be delegated to AI, each grounded in a subject-specific or relational reason. A history teacher's reasons must sound like history.
Draft · 1,285 ch Submitted
04
Mental Model Audit
Annotate 3 AI education marketing texts: mark accurate claims vs. overstatements about AI's nature or teacher agency. Cite specific phrases.
Draft · 1,203 ch Submitted
Day 2 Designing and Evaluating Extensions D2 · Extension Operation Capability
05
Prompt Engineering Log
≥8 annotated iterations. For each: state the pedagogical intent, full prompt, AI output summary, evaluation against the intent, and explicit revision rationale before the next attempt.
Draft · 2,976 ch Submitted
06
Three Output Evaluation Reports
Identify ≥1 hallucination, bias, or over-confident claim per AI-generated material. Explain why the error matters in this subject; write a corrected version. At least one evaluation must address a subtle error.
Draft · 1,828 ch Submitted
07
Customised Extension Artefact
Prompt template, system prompt, or task chain for a B- or C-type activity. Include pedagogical rationale, target HAA type, boundary conditions, and ≥1 tested failure mode.
Draft · 632 ch Submitted
08
Tool Selection Rationale
Compare two AI tools for one unit activity using the A/B/C taxonomy. Evidence of direct testing of both tools required.
Draft · 718 ch Submitted
Day 3 Teaching with Extensions D3 · Extension-Integrated Teaching Practice
09
25–30 min Video-Recorded Micro-Lesson
Implement ≥2 different HAA extension types. Screen visible when AI tools are used. ≥1 explicit mediation moment timestamped in the decision log.
Draft · 673 ch Submitted
10
Student Extension Guidance Episode
≥8 min within the lesson. Teach students how to think about AI as a cognitive extension in this subject. Address: extension type in use, subject-specific boundary conditions, ≥1 concrete override case.
Draft · 1,038 ch Submitted
11
Live Decision Log
≥8 timestamped entries. For each AI use, adjustment, override, or deliberate exclusion — record extension type, pedagogical rationale, retrospective assessment. ≥1 entry must record a deliberate exclusion with a subject-specific reason.
Draft · 1,557 ch Submitted
12
3 Student Work Sample Evaluations
Apply the Output Evaluation framework to AI-assisted student work. Identify AI errors/biases, name likely misconceptions, write feedback addressing the student's AI use — not just their work product.
Draft · 1,790 ch Submitted
Day 4 Learning about Learning D4 · Extension-Enabled Professional Development
13
AI-Assisted Self-Analysis
≥600 words. Upload Phase 1–3 artefacts to an AI assistant; ask it to identify patterns. Annotate every insight as Accepted/Modified/Rejected with subject-specific reasoning. ≥2 rejections must identify an echo-chamber effect.
Draft · 1,163 ch Submitted
14
New AI Tool Evaluation Report
Evaluate an assigned tool not previously used. ≤45 min, timed. Apply A/B/C taxonomy + boundary analysis. State extension type(s), boundary conditions, pedagogical potential, primary risk, adoption recommendation.
Draft · 555 ch Submitted
15
Three Stakeholder Communications
Student, parent, administrator. Each must specify: extension type, decision rationale, AI's limits. Language must be demonstrably different across audiences.
Draft · 1,906 ch Submitted
16
3-Month PD Plan
Code each goal as AI-proposed / self-identified / modified-from-AI. Include a verification method for each. ≥1 goal addresses a specific Phase 1–3 limitation.
Draft · 823 ch Submitted
Day 5 Auditing the Extension D5 · Extension Ethical Responsibility
17
Three-Layer Ethical Audit
Purpose ethics, process ethics, structural ethics. ≥2 risks per layer, each naming the exact lesson activity, affected student population, and mechanism by which harm could occur.
Draft · 1,736 ch Submitted
18
Ethical Risk Memo
800–1100 words. For each risk: Risk name → Mechanism → Affected students → Mitigation action with timeline → Whether revised into the unit plan or accepted with documented rationale.
Draft · 1,236 ch Submitted
19
Purpose Drift Audit
Review every AI use decision in Phases 1–4. For each efficiency-driven decision: either revise or provide a principled argument.
Draft · 1,475 ch Submitted
20
Revised Unit Plan + Revision Log
For each change: cite the audit finding, state what changed, explain expected effect on student development. ≥1 revision must change tool/platform — not just use behaviour.
Draft · 1,106 ch Submitted
Day 6 Anchoring What One Cannot Anchor Alone D6 · Collective Agentic Practice
21
Ambient-Extension (Type F) Audit Memo
500–800 words. Pick ONE ambient (Type F) AI feature in your school's platforms — LMS content recommender, adaptive sequencing, notification/nudge system, analytics dashboard that frames student identity. Name the mechanism, the affected student population, what it is silently doing to student cognition, and why it is invisible by design.
Draft · 534 ch Submitted No AI access
22
Professional-Community Evaluation
Co-author a shared evaluation of ONE AI tool with ≥1 colleague (in-person or async, same subject or cross-subject). The artefact must include: a recorded disagreement, how it was resolved (or why it remains open), and one insight that neither of you had alone.
Draft · 510 ch Submitted
23
Institutional Escalation Draft
Written escalation to leadership (HoD, VP, curriculum lead, tech coordinator) about ONE AI-related risk at your institution that individual teachers cannot mitigate. Must specify: the risk, the mechanism, affected parties, the concrete ask, the timeline, and the evidence base.
Draft · 439 ch Submitted
24
Policy-Level Response
Respond substantively to ONE live or recent AI-in-education policy item — a school district consultation, a national curriculum comment period, an internal PD policy, a professional-association position paper. ≥300 words. Must take a position, not summarise.
Draft · 457 ch Submitted
Day 7 Who Gets to Anchor D7 · Equity-Centred Extension Practice
25
Differential-Effect Map
For each major AI use in your unit, map its differential effect across ≥3 named learner groups in your actual class. Each row: AI use, learner group (EAL / SEN / neurodivergent / culturally under-represented / socioeconomically disadvantaged / etc.), the differential effect, and the evidence you base it on.
Draft · 839 ch Submitted
26
Inclusive Redesign
Take ONE row from p7t1 where you identified a disadvantaged group, and redesign the AI activity so that group is no longer structurally disadvantaged. Document the original, the redesign, and a theory of change linking the two.
Draft · 480 ch Submitted
27
Training-Data Representation Audit
Pick ONE AI output in your subject (a generated example, explanation, or scenario) and audit it for representational gaps. Document ≥3 specific gaps — whose contexts/dialects/histories are present or absent — and trace each to a pedagogical consequence in your unit.
Draft · 959 ch Submitted
28
Anchoring-Capacity Cultivation Plan
Plan how you will cultivate anchoring capacity in learners who came to your class with less AI exposure or lower confidence exercising agency over AI. Name specific learners (initials only), specific interventions, and how you will know it's working.
Draft · 990 ch Submitted
Day 8 Rehabilitating Assessment Validity D8 · Assessment Redesign Under AI
29
Assessment Validity Audit
≥2 existing assessments (yours or departmental). For each: what was it designed to measure? what can AI now produce that would look identical to that capacity being exercised? what is the failure mechanism that makes the assessment no longer a valid signal?
Draft · 813 ch Submitted
30
Authentic Agentic Assessment Design
Design ONE new assessment that is STRUCTURALLY non-evidential for AI-substituted output — oral examination, in-class writing, process portfolio, scaffolded real-time construction, iterative review with explained reasoning, or think-aloud protocol. Specify: construct measured, form, why AI substitution fails to provide valid evidence for this form, scoring approach.
Draft · 431 ch Submitted
31
AI-Inclusive Policy Co-articulation
Run ONE conversation with your class (or a focal group ≥4 students) about legitimate vs. illegitimate AI use in your subject. Record what they named as legitimate / illegitimate / ambiguous. Produce a short shared policy document.
Draft · 519 ch Submitted
32
Revalidation Trial
Run p8t2's new assessment with ≥5 students. Write up what the evidence actually showed: did the form discriminate agentic work from AI-substituted output? what surprised you? what revision do you propose?
Draft · 508 ch Submitted

Teacher Context

Large class, limited lab equipment. Following CAPS curriculum. Some students have weak science backgrounds from previous grades.