The template
| Field | What goes here |
|---|---|
| Tool(s) used | Name and version of every AI tool you used. Example: "Gemini 2.5 Flash (May 2026)"; "Claude Sonnet 4.6"; "Plate Tectonics Tutor (BoodleBox box)". List all if you used more than one. |
| Prompts (representative) | 1–3 of the prompts that shaped the work most. Quote them. You don't need every prompt: just the ones that mattered. |
| What the AI contributed | Specific tasks: brainstorming, outlining, drafting a paragraph, giving feedback, translating, summarizing a source, explaining a concept. Be specific: "it helped me" is not enough. |
| What YOU contributed | Decisions, evidence you chose, your interpretation, your voice, your edits, your conclusions. This should be the bulk of the work. Then name one part you deliberately kept for yourself, and say why that part. |
| OPVL of the AI's contribution | A short OPVL on the AI as a source for this specific work. Highlight especially the Limitations: what was it wrong about, what did you fix, what couldn't it do? See OPVL Framework for a refresher. |
| One thing the AI got wrong | Required. Name at least one specific thing the AI got wrong, missed, refused, or couldn't do well. If you couldn't find one, that's a sign you weren't checking carefully enough. |
| Was it worth it? | One line. Running an AI costs energy and water. Was what you got back worth that cost, for this task? A reasoned no is as good an answer as a reasoned yes. |
Guiding questions for each field
The table says what goes in each field. These are the questions to actually answer while you write it. They are deliberately general, so they work for every block of the year, whichever AI you used and whatever the task was.
- Tools used. Which AI tools did you use, and which version? If you used more than one, which did what?
- Prompts. Which one to three prompts changed your work the most? Quote them exactly, not from memory.
- What the AI contributed. What did the AI actually do? Name the task, not the feeling. "Outlined three counter-arguments" is a task. "It helped me" is not.
- What you contributed. Which decisions, evidence, and conclusions are yours? Which part did you deliberately keep for yourself, and why that part?
- OPVL. Where did this output come from, what was the tool built to do, what was it genuinely useful for here, and what could it not do?
- One thing it got wrong. What is one specific thing it got wrong, missed, refused, or handled badly? How do you know? What did you check it against?
- Was it worth it. Running the model cost energy and water. For this task, was what you got back worth that? A reasoned no scores as well as a reasoned yes.
Question 6, and specifically the second half of it. Anyone can assert that an AI got something wrong. Naming what you checked it against is the part that shows you actually checked. A source, a page number, a search you ran, a passage you reread: that is the evidence, and it is what the higher bands are looking for.
Watch for the trap: a vague response is not a more reliable one, it is just harder to catch. If a response was too general to verify, say that. "Too unspecific to check" is a real Limitation and it belongs here.
If this template is missing or trivially completed, the work is returned as incomplete and is not scored until the documentation arrives. No strand is capped.
The "one thing it got wrong" field is part of a complete template, and it is the metacognitive move T/T is built to detect. A specific answer there is what earns the higher bands.
Need to understand why each field exists?
The AI Documentation Protocol in Foundations explains the reasoning behind each field, lists the AI failure modes to watch for, and walks through three worked examples (strong, weak, and the violation). Read it once at the start of the year; come back when you're unsure what counts as "specific enough."
Mini-version: for incidental AI use only
For very minor AI use (one prompt, a task you can already check, e.g. "Gemini translated one paragraph from Hindi"), use the three-field mini-version:
- Tool + one prompt
- What I did with it
- One thing it got wrong (or that I had to fix)
Default to the full template for graded work. The mini is for incidental use that doesn't shape the deliverable. If in doubt, use the full template.
If you're unsure whether your use needs documentation
"Could I explain to my teacher exactly what I did, and would I be comfortable doing so?"
If yes, document it and submit. If no, stop and recalibrate. See Class Expectations for the full disclosure principle.