How we use AI
Proposal for the teachers' meeting · 2 October 2026 · Built on Heinz's AI policy draft (May 2026)
The whole process in 70 seconds
Sound on · subtitles in the player menu
Built on Heinz's draft
“If you cannot explain a line of code to your teacher or a teammate, it does not belong in your project.”The Golden Rule · AI Use Policy Addendum · Heinz K, TECHCOLLEGE · May 2026
Three things every student needs to know
Any AI, any part
- Any tool, including agents that build whole features while you direct and review
- One exception: design sprint interviews are with real people
Notes, every day
- A few bullet points: what the AI got wrong, what you changed, what you learned
- The AI turns each day into an interactive HTML page with your notes and its own account
- A team AI report built from both
Two rubric items
- Understanding your code: each student explains their own code to a teacher
- The AI report: each team, after the project
- How critically, not how much
Every day: your notes + the AI's own account
- Try it
- The problem
- Decision log
- What the AI got wrong
- Tutorials
- How we used AI
One student's project days, from the first commit to the graded report.
The AI report: report/index.html
21 January: group visits, not presentations
Three mixed pairs of teachers visit eight groups.
6 teachers, two from each school. Every pair mixes two countries.
groups per pair, about 8 groups in all.
per pair: 15 min of group questions plus 3–5 min per student.
Two new rubric items
| 0–2 | 3–4 | 5–6 | 7–8 | 9–10 | |
|---|---|---|---|---|---|
| 02 | 4 | 7 | 10 | 12 | |
| Understanding your codeeach student | Cannot explain own code | Explains own code only with help | Explains what, but not why | Explains why, answers follow-ups | Weighs alternatives, changes it on the spot |
| The AI reporteach team | No report or notes | Notes written at the end | All sections, some AI mistakes | Daily notes tied to commits, clear decisions | Catches subtle mistakes, teaches others |
Why notes and a conversation, not a declaration
of students didn't declare their AI use, even with a mandatory coversheet.
Gonsalves, King's College London, 2025of student repos named an AI tool, but almost none said what it did.
Parreira et al., 203 repositories, 2026of essays by non-native English writers were wrongly flagged by AI detectors.
Liang et al., Patterns, 2023Rules that only tell students what they may do create an “enforcement illusion”. What works is changing the assessment itself. Corbin, Dawson & Liu, 2025
What we decide today
Use AI.Own every line.Show your work.
Thank you. The full proposal, the policy text and the film are on the page.
ellertsmari.github.io/AI-report-suggestion/