Is AI stopping your students from thinking hard?
It’s a fundamental principle of cognitive science: hard thinking is critical for long-term learning. A core tenant of “desirable difficulties” highlights how if students aren’t engaged in effortful thought, there is a good chance that they aren’t doing much learning.
AI, however, can easily bypass this process. It can write essays at the click of a button, doing all the hard thinking for our students. While this might seem efficient, it undermines long-term learning. If students get used to AI doing all the thinking for them, what happens when they get in an exam, and they no longer have access to AI?
In essence, how can we use this technology in a way that promotes—not detracts from—hard thinking? In order to do this, I have tried to develop strategies where I use the AI to support students to think hard, thereby modelling effective use.
Here’s what I’ve tried:
Using AI to Identify Contemporary Examples
One challenge students often face is coming up with relevant, contemporary examples for their essays. For instance, in Sociology a common critique of some definitions of globalisation is their tendency to focus exclusively on social, political, or economic factors without recognising how they intersect. I used ChatGPT to provide examples that demonstrate this intersection:
- COVID-19: Social inequalities were exposed, lockdowns affected social connections, international cooperation highlighted political factors, and global trade was economically disrupted
- Facebook: A platform that connects people socially, generates economic revenue through advertising, and plays a political role in movements like #BlackLivesMatter or election campaigns
Once these examples were identified, students used them to construct their arguments. The key, however, was ensuring they didn’t just copy and paste. Instead, they rewrote the examples in their own words and critically analysed how they could support or challenge sociological perspectives. This engaged them in deeper thought, ensuring they weren’t just using AI passively but working with the material in an active way.
Using AI to Create Error-Ridden Paragraphs
Another strategy was using AI to write a deliberately flawed paragraph. Students were tasked with identifying and correcting the errors, which required close reading and critical engagement. Their final product was an opening essay paragraph that reflected their active engagement with the material.
By analysing and improving the AI-generated paragraph, students thought hard about what makes a response accurate, coherent, and effective. This was real time saver for me, I could have written the paragraph myself but instead I put a few bullet points into ChatGPT, uploaded a PowerPoint from my lesson with the correct content to include, and asked for a paragraph summarising the theory.
I then asked for the paragraph to be edited with deliberate errors that the students could correct. One of my top tips when using AI to construct paragraphs is to provide it with examples of the kind of thing you want, or with the content to include. The more information you can give about what you want e.g. include the exam board and the age of students, the more likely the output is going to be useful.
Using AI for Comparative Examples
AI is also a great tool for creating comparative examples of written work. I used ChatGPT to generate two contrasting paragraphs, each illustrating different levels of quality in writing technique. Students then analysed and compared them, identifying what made one response stronger than the other. From this analysis, they extracted techniques to apply to their own work.
Here, too, the AI served as a scaffold—not a shortcut. It provided models that supported student learning, but the hard thinking came from evaluating and applying those models. I have even asked AI to generate step by step instructions of how to write a fully developed paragraph using the models I provided it with and it immediately broke the task down into small achievable steps for my students.
The Importance of Active Engagement
The temptation to let AI do all the work is real, but this approach fails to promote meaningful learning. If students simply click a button, copy a response, and submit it without engaging with the material, they miss out on the deep thinking required for long-term retention. Instead, by teaching students the principles of cognitive science and giving them structured ways to use AI as a tool—not a crutch—we can ensure they use it in ways that foster real understanding.
A Final Thought: AI as a Time-Saver
AI is incredible, not just for students but for teachers too. Full disclosure: this blog itself was supported by AI, the first draft crafted from my hurried collection of bullet points. With thoughtful application, AI can be both a time-saver and a powerful tool for promoting hard thinking and long-term learning.
This blog was written by Rachel Davis, an associate assistant headteacher. To learn more about helping your staff to use AI tools effectively, check out our AI in education workshop.
