Pilot the AI tool on your best-known topic
Before any tool reaches students, run it on the topic you know best. That is the one place you can see exactly what it gets wrong.
Nov 9, 2026 · 1 min read
You cannot judge an AI tool's reliability on a topic you half know; its errors just look like information. So pilot every new tool on your strongest unit, where you can spot distortion instantly. Note what it represents well, what it flattens, and whether students could complete your tasks with it while skipping the thinking those tasks exist to develop. Mishra and Koehler (2006) argue that sound technology integration lives at exactly this intersection of tool, subject and pedagogy. (Manning's technological-pedagogical-content-knowledge-developer skill walks the full review.)
In practice
I teach [TOPIC] to [YEAR GROUP].
Answer the five questions below exactly as you
would for one of my students:
[PASTE YOUR STUDENTS' FIVE COMMONEST QUESTIONS]
Then explain [THE CONCEPT THEY FIND HARDEST]
step by step.
(Now compare every answer against what you
know cold. Note distortions, gaps and anything
a novice would swallow whole.)- Ask it the five questions your students always ask, and compare its answers with the ones you have refined over years of teaching.
- List what the tool makes visible and what it obscures; a graphing tool shows shape but hides the computation underneath.
- Name the dependency risk in writing: which thinking step can students now skip, and does your task still force that step?
- Decide the walk-away test before first classroom use: what evidence would make you drop this tool mid-term?
Quick takeaway: audit the tool where your expertise is deepest, because that is where its errors are visible.