Teaching students AI literacy
Whether students meet AI tools no longer depends on the school’s permission. They already run into them on their phones, on the computer at home and inside the apps they use every day. That leaves schools with two options: ignore this use, or help students use these tools with understanding and a critical eye. I think the second is both more realistic and more educational.
When people hear “AI literacy,” many think of knowing how to use a tool well, in other words writing effective prompts. That is part of it, but the smallest part. What really matters is that students understand what these tools are, how they work, where they can and cannot be trusted, and when using them will harm their own learning. As with media literacy, the aim is less about teaching a tool and more about making students’ relationship with it a conscious one.
The first step is to take the mystery out of the tool. Many students perceive chat-based AI as something that knows everything, thinks and even feels. Correcting this in age-appropriate language is important. Even in middle school you can explain the core idea: these tools were trained on enormous amounts of text and produce answers by predicting the words most likely to follow a question. They write something not because they know it, but because that kind of answer is likely to come next.
Rather than leaving this as an abstract explanation, you can make it concrete with a small classroom activity. Write half a sentence on the board and ask students to guess how they would finish it. You will see that most of them choose similar words. Then talk about how these tools make similar predictions, only on a far larger scale. A simple activity like this makes it easier for students to find their own answer to “why does AI sometimes get things wrong?”
Seeing the mistakes together in class
Perhaps the most important part of AI literacy is for students to see with their own eyes that these tools can give wrong information in fluent and confident language. Telling them is not enough; they need to experience it. Pick a topic your students know very well. A detailed question about the place where your school is, a fine point from a topic you have just covered, or a local event all work well. Look at the tool’s answer together and ask students to find the errors.
When students catch the mistake themselves, the awareness that follows is strong. You are also practising verification. How do we know whether a piece of information is true, which source do we check, what do we do when two sources disagree: these questions matter more than ever in the age of AI. In that sense AI literacy is a new extension of the information literacy we have been trying to teach for years.
Sources that do not exist deserve their own conversation. These tools sometimes cite books, articles or web pages that are not real. With high school students, looking up the sources a tool suggests one by one and checking whether they actually exist shows this risk in a way they will not forget.
Bias is another topic worth covering. These tools reproduce patterns from the texts they were trained on, and in those texts some perspectives, languages and cultures are far better represented than others. Having the tool write descriptions of a profession or a country and then discussing with students what assumptions those texts contain turns an abstract idea into something very concrete.
When not to use it
While teaching students how to use AI, we also need to teach them when not to. If the purpose of an assignment is to develop a student’s ability to summarise a text, having a tool write the summary completes the assignment but not the learning. Explaining this as a fact about learning rather than as a rule works better. You can compare it to going to the gym and having someone else lift the weights for you. The weights go up, but your muscles do not grow.
That is why I recommend setting clear rules about AI use in assignments. Students should know from the start in which assignments it is allowed, in which it is not, and how to declare it when they do use it. Ambiguity pushes students to hide their use, while clear rules make honest use possible. A simple approach is to sort assignments into three groups:
- Assignments to be done without AI
- Assignments where AI can be used at specific stages, such as generating ideas or getting feedback
- Assignments where working with AI is part of the task itself
The third group is especially valuable. For example, you can give students a text the tool wrote on a topic and ask them to find its errors, fill its gaps and rewrite it with their own reasoning. In an assignment like this, the student uses the tool, but the thinking load is still on them.
Asking for a short reflection note on assignments where AI was used also helps. A student who writes a few sentences about at which stage they used the tool, what they asked, which part of the answer they accepted and why, and what they changed starts to notice their own habits. These notes also give you a valuable picture of how your class is actually working with these tools.
Privacy is an inseparable part of AI literacy too. Students often do not think about what it means to type their personal information, photos of their friends or private problems into these tools. Knowing that anything they write may be stored somewhere and used for other purposes is similar to the caution we already teach for social media.
There is an emotional side as well. Some students hold long conversations with chat-based tools as if talking to a friend. A tool that is always polite, always available and always agreeable can be appealing, especially for students who feel lonely. This needs to be discussed without judgement, with a reminder that the tool is not a person and cannot replace real relationships.
AI literacy does not have to be a separate subject. Language classes can work on evaluating texts, science on evidence and verification, social studies on source criticism and computing classes on how these systems work, so every subject does its share. What matters is that the school approaches the topic with a shared language and students do not face a different set of rules in every class.
It is almost certain that our students will work alongside these tools in the years ahead. The most valuable thing we can give them is the habit of not switching off their own thinking while they use them.
See you soon.
Image: Office of the President of Ukraine, CC0, Wikimedia Commons.