On this page
It is tempting to finish a difficult physics or chemistry chapter, drop the text into an AI, and ask for a quick summary. However, asking an AI to do the summarising for you skips the exact cognitive effort that your brain needs to learn. If you write your own summary first and then compare it to the AI's version, you can spot exactly what you misunderstood while still building long-term memory.
When you read a perfectly structured AI summary, it feels like you understand the material. Psychologists call this the illusion of competence. Because the text flows logically, your brain recognises the information and assumes it is stored safely. But recognition is not the same as recall. If you sit down for the exam the next day, you will likely struggle to pull those neat details from memory because you never practised retrieving them.
Why does writing a summary improve memory?
The alternative is to write the summary yourself. This feels harder, and it takes longer, but that difficulty is exactly what builds robust memory.
When you create your own version of a text, you trigger the generation effect. A meta-analysis of 86 studies (Bertsch et al., 2007) found that actively generating information rather than passively reading it leads to a reliable memory advantage, with an average effect size of 0.40.
Writing a summary forces you to make active decisions. You have to retrieve the facts from memory, decide which minor details to drop, and connect the main ideas in your own words. This cognitive effort creates strong neural pathways, converting what you just read into what you actually know. It is a prime example of desirable difficulties in practice: the harder it feels to remember the material now, the easier it will be to recall it during the test.
How do you compare your summary with AI?
Instead of starting your study session with AI, use it as an expert reviewer after you have done the heavy lifting. This framework takes an extra ten minutes but locks in the chapter.
Read and close: Read the chapter or your class notes, then close the book. You must write from memory, without glancing back at the source text. This step is non-negotiable; if you look at the book while summarising, you are just copying, and the memory benefits disappear.
Draft your summary: Write down the core ideas in your own words. Aim for five sentences or a short mind map. Do not worry about making it sound perfect; focus on capturing the physics or biology accurately. If you get stuck on a detail, leave a blank space and keep moving.
Ask the AI: Now, bring the topic to Fermi's AI tutor and ask it to generate an overview of the exact same subject.
Compare the two: Read both summaries side-by-side.
What should you look for when comparing summaries?
The differences between your rough summary and the AI's polished one are a map of your weak spots. You are not looking for style; you are looking for missing concepts and broken links.
For example, imagine you are revising Newton's laws of motion. Your summary might list the three laws correctly but fail to mention that they only apply in inertial frames of reference. When the AI includes that crucial condition, the gap tells you exactly what to revise immediately.
If the AI's summary uses a key technical term that you struggled to recall, that is a signal to review the underlying vocabulary. The goal is not to write a summary as perfectly polished as the AI's. The goal is to use the comparison as an instant feedback loop. You get the benefit of tutoring hints tailored specifically to where your own memory stopped.
You can even take it a step further by showing your summary to the AI and asking, "What did I miss?" The tutor can then point out your specific conceptual gaps. By writing first, you guarantee the learning happens in your brain. By comparing second, you guarantee you haven't missed the point.
