Unlocking The Power of AI: How It Deepened My Students’ Thinking

Chantelle Love
4 min readJan 31, 2023
Image created with DALL:E2

Something I’ve encouraged my students to do in recent years is to develop their complex problem solving by creating ‘ungoogleable’ questions… questions that are so complex one can’t simply ‘google’ it. Students respond in different ways to this challenge but what’s even more confronting is what happens after that.

If we can’t google the answer, how will we do our research?

That makes things awkward. And, thus, students have resisted because research began to become cumbersome, often resulting in the point of the research getting lost.

With the recent surge in AI, I revived ungoogleable questions in our Year 5 classroom for a little experiment — would AI have the capability to facilitate ungoogleable questions and thus build deeper learning about the topic (as noted by Wu & Chen, 2018)?

“AI algorithms have the power to provide users with a more in-depth understanding of their topic of interest.”

So, how did we start?

  1. Build familiarity with prompt engineering: We began by upskilling students in prompt engineering. For EAL learners, this was a game-changer as it gave language to specific commands. I created a less sophisticated version of this article for my students to use.
  2. Use concepts and content to create questions: Being a PYP school, using conceptual thinking is our bread and butter.
Image from Structural Learning

Knowing that our students had already chosen their “content” (a migrant’s story), we placed A3 posters around the learning space with each concept written in the middle. Over 10 minutes, students silently wandered to each poster and wrote questions mashing their content and the concept together (teachers were also involved). They then spent the next 10 minutes reading the posters and choosing at least one conceptual question from each poster to guide their research.

Some of the questions students developed.

NB. Part of this process was also the critical thinking involved in choosing questions that weren’t suitable for their content.

3. Model how the AI works: we used perplexity.ai for our non-googleable questions.

We chose perplexity.ai because students don’t require a sign-in and also, it provides sources for the responses. I encouraged my students to read the ‘detailed’ response as it was often more accurate.

Our first discussion point was, ‘Did my question get answered?’ As you can see, it was only half answered so we can go back and change our prompt. Our second discussion point was, ‘Was the response accurate?’ and together, we interrogated the sources.

“AI algorithms have the ability to identify and filter out low-quality or unreliable information, ensuring that users receive accurate and trustworthy information.” (Liu, 2018)

4. Ensure that the assessment is ungoogleable and un-AI-able. We decided that the assessment would be a museum box about the subject with a journal/blog/letter written in a moment in time, a map showing the migration, a timeline and several artifacts. Students would share their learning in a mini-expo format to parents and other students.

And… did students have a more in-depth understanding of their topic of interest?

From what I could tell, and it hasn’t been peer-reviewed or anything, but students did. They were incredibly confident talking about their subject and they were incredibly knowledgeable about their subject. For example, they knew that Einstein had concerning articles written about him by the Nazi’s and thus why we fled so quickly. They knew Anne frank’s favourite colours. They knew Tesla’s patent information. I’ve never seen student’s uncover that depth of information before.

I also believe they were able to transfer that information more effectively as well.

The Verdict

Overall, I was super-impressed with the way AI enhanced my students’ depth of thinking.

I’d encourage you to try using AI with your students and see if you agree.