Chapter 2. What AI Is, and What It Is Not
Teachers who watch PakEdX at work for the first time often ask me the same question. A lesson plan fills the screen in seconds, neat and confident, and someone leans forward. ‘Where does it find all this?’ Most expect me to name a website, or a vast library of lesson plans sitting somewhere on a server. The honest answer surprises them. Underneath, the language model finds nothing at all. Unless someone connects it to a search engine or a set of documents, it simply writes the most likely next words, one small piece at a time. It learned the patterns behind those words long before the lesson began.
Teachers have good reason to ask. In my doctoral research, one university teacher said students ‘are using AI a lot’ and ‘often know more than I do’ about using technology (Ilyas, 2026, p. 161, Participant M01). Other teachers in the study said much the same. I described this in my thesis as a role reversal, with students becoming a resource for teachers’ own learning (Ilyas, 2026, p. 161). A role reversal is easier to handle when teachers understand the tools their students already use.
In this chapter
This chapter asks what teachers need to understand about how AI works. The answer is less technical than many fear. A teacher does not need to build a model, any more than a driver needs to build an engine. Yet a basic understanding of what happens inside the tool is enough to change how we check, trust and teach with these tools. The chapter builds that picture in four steps, looks at where teachers stand today, then turns the picture into ground rules for safe use.
What the chapter covers
- What happens when you press enter
- Fluent is not the same as true
- Why it makes things up
- Where bias comes from
- Where teachers stand
- Ground rules for safe use
- In practice
- When resources are thin
- Questions for reflection
- Framework connections
Questions for reflection
Take these questions to your next department meeting, or into a quiet moment after school.
- When your students use a chatbot, what do you think they imagine it is doing behind the screen?
- Which topics in your subject sit where training data are thin, and how would you check answers there?
- Have you ever asked a tool ‘Are you sure?’, or told it you wrote something, and what did its reply really tell you?
- Where in your school does personal data meet AI tools, and who approves those tools?
- Which language do your students think in, and how well do the tools serve it?