Module 1 Understanding AI in Education
A non-technical foundation. Teachers leave able to tell rule-based software, machine learning and generative AI apart with classroom examples, to recognise a confident wrong answer, and to place AI correctly in a teaching decision: it analyses and suggests; the teacher diagnoses and decides.
By the end, teachers can…
- 1 Explain, in non-technical terms, the difference between rule-based software, machine learning and generative AI.
- 2 Give two classroom examples each of what current AI does reliably and what it does unreliably.
- 3 Recognise a hallucination and describe how to verify AI output before using it with students.
- 4 Describe human-in-the-loop decision-making and apply it to one routine teaching decision.
- 5 State the difference between a prediction about a student and an understanding of that student.
Session plan — 2 hours
| Time | Activity | Format |
|---|---|---|
| 0:00 10 min | Opening: what do you already use? Pairs list every AI-touched tool they used this week — maps, autocorrect, recommendations, a chat assistant. Surfaces that "AI" is already in the room and demystifies the word. | Discussion |
| 0:10 20 min | Three kinds of "AI" Rules vs learned patterns vs generative models, one school example each: a timetable program, a spam filter, a chatbot. The through-line: prediction, not understanding. | Input |
| 0:30 20 min | Reliable / unreliable sort Groups sort sixteen claim cards ("summarise this chapter", "predict which student will fail", "grade this essay fairly", "explain why a student got it wrong") into reliable today / sometimes / not yet — and justify each. | Group work |
| 0:50 25 min | Hallucination lab Teachers ask any chat assistant three questions from their own subject — one factual, one about a recent board circular, one asking for a citation — and check each against a source. They record what was wrong and how they found out. | Hands-on |
| 1:15 20 min | Human in the loop Where AI sits in a teaching decision. AI analyses, suggests, personalises, practises, tracks, assists. The teacher diagnoses, guides, challenges, contextualises, motivates, validates. Three short cases where a teacher rightly overrode an AI suggestion. | Input |
| 1:35 25 min | Reflection + foundations check Individually: "one thing I will now verify before using; one decision I will never hand to AI." Then a ten-item scenario quiz (counts towards Level 1). | Check |
| Total 2 h · timings are a default; facilitators adapt to the group. | ||
Key ideas
Prediction, not understanding
A model that says a student "is likely to struggle with fractions" has matched a pattern in data. It has not met the child. The teacher has — and that is why the teacher decides.
Confident is not correct
Generative systems produce fluent text whether or not it is true. Fluency is not evidence. The professional habit is: verify before use, especially for facts, citations and anything about a named student.
Augmentation, not replacement
The useful question is never "can AI do my job?" but "which part of my week is pattern-work that a machine can do first, so that I can spend the time on the part only I can do?"
Accountability does not transfer
If an AI suggestion reaches a student or a parent, the teacher owns it. That is not a burden added by AI — it is the same professional standard, applied to a new tool.
Worked example
Reading an AI "insight" about a student
A dashboard flags: "Aarav — engagement dropping; score slipping on Linear Equations."
Prediction: the pattern in Aarav’s recent attempts looks like the pattern of students who fell behind. Useful — a reason to look.
What the teacher knows: Aarav missed four days for a family event; his last two attempts were done on a phone at night.
Decision: a two-minute check-in tomorrow, not a remediation plan. The flag was right to ask; the teacher was right to decide. That sequence — AI recommends, teacher decides — is the whole module in one example.
Assignment — counts towards certification
Verify-before-use log
Over one week, use any AI assistant for three real teaching tasks (e.g., draft a worksheet, explain a concept two ways, summarise a chapter). For each, record what you asked, what you kept, what you changed, and what you had to verify.
Deliverable: A one-page log (three entries) plus a three-sentence reflection on where AI helped and where it misled.
| Criterion | Meets the standard when… |
|---|---|
| Three genuine tasks | Each entry is a task the teacher actually needed that week, not an invented prompt. |
| Verification shown | At least one entry shows a factual check against a source, with the source named. |
| Judgement exercised | At least one entry shows something kept, something changed, and why. |
| Reflection | Names one reliable use and one unreliable use in the teacher’s own subject. |
Where this lands in the PrepGraph workflow
Each certification task corresponds to a behaviour in the teacher portal — so the programme produces a working habit, not a slide deck. The pedagogy comes first; the surface implements it.
| Certification task | Pedagogical move | PrepGraph surface |
|---|---|---|
| Read an AI signal about a student critically before acting on it | Human-in-the-loop review | Early-warning signalsNightly signals per group — engagement dropping, scores slipping — so you know who needs a check-in this week. |
Materials
- •Sixteen "claim cards" (printable set in the facilitator pack)
- •Any chat assistant on a phone or laptop, one per pair
- •Foundations check (ten scenario items, answer key)
- •Slides: three kinds of AI; human in the loop
Facilitator notes
- •Keep it non-technical: no architectures, no vendor comparisons. If asked "which AI is best?", redirect to "best for what task, verified how?"
- •In the hallucination lab, make sure at least one pair asks about a CBSE circular by number — models routinely invent plausible circular text. That moment teaches more than any slide.
- •Do not demonstrate PrepGraph in this module. The only product touchpoint is the worked example, read as a story.
- •Expect and welcome scepticism; the module succeeds when a sceptical teacher says "so it’s my call" — that is the principle.
Bring the programme to your school
Tell us your board, grades and teacher count. We’ll reply within two working days with a cohort plan and the CPD record templates.