Study prompts designed around what actually produces recall: retrieval practice, spacing and explaining things in your own words.
5 copy-ready prompts
Retrieval focused
Any subject
The prompts
Prompt 1
Explain it at three levels
Meeting a new concept for the first time.
Explain [concept] three times:
1. To someone with no background at all, using an everyday analogy.
2. To a [your level] student, with the correct terminology.
3. To someone who will be examined on it, including the details that are commonly tested and the mistakes people usually make.
Then tell me where the analogy in version 1 breaks down, so I do not carry a wrong mental model forward.
Why it works: The last instruction matters. Analogies teach quickly and mislead quietly - knowing the edge of one keeps it useful.
Prompt 2
Test me instead of telling me
Active recall, the single most effective study technique.
You are quizzing me on [topic] at [level].
Ask me one question at a time. Wait for my answer before continuing.
After each answer:
- Tell me if it is right, partly right or wrong.
- If it is partly right, say exactly which part was missing.
- Do not give me the full answer until I have attempted it.
Start with recall questions, then move to application questions. Stop after 10 questions and tell me which areas were weakest.
Why it works: Retrieval feels harder than rereading, which is why people avoid it. That difficulty is the mechanism.
Prompt 3
Find the holes in my understanding
Checking whether you actually know something.
I am going to explain [topic] in my own words. Read it and tell me:
1. What I got right.
2. What is wrong.
3. What is missing that matters.
4. Anything I stated confidently that is actually more complicated.
5. Two questions that would expose whether I understand this or have just memorised it.
My explanation:
"""
[write it out without looking at your notes]
"""
Why it works: Write your explanation from memory first. Looking at your notes turns this into copying rather than testing.
Prompt 4
Build a spaced revision schedule
Planning backwards from an exam date.
I have an exam on [date] covering [topics]. Today is [date]. I can study [hours] per week.
Build a revision schedule that:
- Revisits each topic at expanding intervals rather than once.
- Puts my weakest topics - [list them] - on the most frequent cycle.
- Includes a retrieval practice session, not a rereading session, for every topic.
- Leaves the last three days for full practice papers and light review only.
Give it to me as a week-by-week table.
Why it works: Schedules that only cover each topic once produce recognition, not recall. Spacing is what moves material into long-term memory.
Prompt 5
Turn a chapter into practice questions
Converting passive reading material into something you can be tested on.
Here is material from [subject]:
"""
[paste your notes or chapter summary]
"""
Create:
- 8 recall questions with short answers.
- 4 application questions requiring me to use the idea in a new situation.
- 2 questions that connect this material to [other topic I have studied].
- 3 plausible wrong answers a student might give, with why each is wrong.
Put the answers at the end, not next to the questions.
Why it works: The wrong answers are a study tool in themselves - misconceptions are easier to avoid once you have seen them named.
Studying with AI without fooling yourself
The easiest way to use AI for study is also the least effective: asking it to summarise everything and reading the summaries. These prompts do the opposite, because the techniques that actually produce recall all feel harder than reading.
Why summaries feel productive and are not
Reading a clear explanation creates a strong sense of understanding. That feeling is fluency, and decades of research on learning show it correlates poorly with what you can produce later without the text in front of you. Highlighting and rereading rank consistently among the least effective study methods studied, despite being the most popular.
Retrieval practice - attempting to produce the answer from memory before checking - is one of the most effective. It is also uncomfortable, which is exactly why students avoid it and why the quizzing prompt on this page is written to withhold the answer until you have tried.
Spacing beats cramming, reliably
Reviewing material at expanding intervals produces far better long-term retention than the same total time spent in one block. The forgetting curve is steep at first and flattens each time you successfully recall something, so a topic revisited after a day, then three days, then a week, then a fortnight, ends up far more durable than one studied intensively the night before.
The scheduling prompt asks for expanding intervals and for retrieval sessions rather than rereading sessions, because a plan built on the wrong activity just spaces out an ineffective method.
Where AI is genuinely well suited
Unlimited practice questions on exactly the material you are studying, which no textbook provides.
Patient re-explanation at a different level or from a different angle, as many times as you need.
Feedback on your own explanation, which is otherwise hard to get without a tutor.
Naming common misconceptions so you can check whether you hold one.
Connecting topics you studied weeks apart, which is where deeper understanding tends to form.
The verification problem
Models state incorrect things fluently, and in a study context you are the least equipped person to notice - you are learning the subject precisely because you cannot yet tell right from wrong. Treat generated material as practice, not as source of truth. Check definitions, formulas, dates and anything that will appear in an exam against your course materials or textbook. If an AI explanation contradicts your lecturer, your lecturer sets the exam.
This is also why the three-levels prompt asks where the analogy breaks down. A simplification that nobody flags as a simplification becomes a misconception you defend confidently in an exam.
A workable weekly pattern
1Learn new material from your primary source - lectures, textbook, course notes.
2Within 24 hours, write an explanation from memory and run the "find the holes" prompt on it.
3Convert the material into practice questions and answer them without notes.
4Revisit older topics on expanding intervals rather than only studying what is new.
5Once a week, mix topics in a single session so you practise choosing an approach, not just applying one.
Notes are for retrieving, not for collecting
Long verbatim notes feel productive and are rarely reread. Notes you can be tested from are a different object: a question on one side and an answer on the other, or a short prompt that forces you to reconstruct an explanation. Converting a lecture into twenty questions takes longer than transcribing it and is worth several times as much, because the conversion itself requires you to decide what matters. It also means every future revision session has material ready, which removes the most common excuse for rereading instead of testing.
Related resources
For technical subjects, the AI Coding Prompts include an explanation prompt built for unfamiliar code, and the Roadmaps section lays out learning order for developer skills. If you are writing essays as part of your course, the ChatGPT Writing Prompts cover drafting and self-critique.
Frequently asked questions
Is using AI to study cheating?
Using it to test yourself, explain concepts and generate practice questions is studying. Submitting generated work as your own is not, and most institutions treat the two very differently.
What is the most effective way to revise?
Retrieval practice spread over time. Attempt to recall material from memory, check it, then revisit at expanding intervals - it consistently outperforms rereading and highlighting.
