What active recall is
Active recall means pulling information out of your memory instead of putting it back in. You close the book and try to answer a question, write down everything you remember about a topic, or explain a step without looking. Rereading, highlighting and watching the lecture again are the opposite, because the information is in front of you the whole time.
Researchers call it retrieval practice, or the testing effect. It's one of the most replicated findings in learning research. In a well-known 2006 study, Henry Roediger and Jeffrey Karpicke had students either reread a passage or practice recalling it. A week later, the recall group remembered more, even though the rereaders had felt more confident.
That last part is the trap. Rereading feels productive because everything looks familiar. Recall feels harder and slower, and that difficulty is what makes it work.
Why ChatGPT helps
The hard part of active recall was never the method. It's writing good questions, which takes time, and most people give up and go back to rereading.
ChatGPT is fast at writing questions, and it can mark a free-text answer instead of only checking a multiple-choice letter. Used well, it's a patient quizzer that never runs out of questions.
Used badly, it's a very fast way to read answers. The default ChatGPT habit is to explain everything at once. You have to tell it not to.
How to make ChatGPT quiz you
Step 1: Give it the material
Paste your notes, a chapter summary or the lecture slides, or upload the file. If you only name a topic, ChatGPT quizzes from what it knows, which may not match your course. Your own material keeps the questions on what you'll actually be tested on.
Step 2: Set the rules before the first question
This prompt does most of the work:
Quiz me on the material above using active recall. Ask one question at a time and wait for my answer. Don't give hints unless I ask. After I answer, tell me what I got right, what I missed, and the correct answer in two or three sentences. Then ask the next question. Start easy and get harder.
"One question at a time" is the most important line. Without it you'll get a list of ten questions with the answers underneath, and your eyes will drift to the answers.
Step 3: Ask for the right kind of questions
Not all questions are equal. In rough order from weakest to strongest:
- Multiple choice. You only have to recognize the answer. Fine for a warm-up.
- Short answer. "What is the function of the mitochondria?" Now you have to produce it.
- Explain. "Explain why the reaction slows down when the temperature drops." You have to connect ideas.
- Apply. "Here's a new case. What would happen and why?" You have to use the idea somewhere you haven't seen it.
- Compare. "How is mitosis different from meiosis, and why does it matter?" This forces you to tell apart things you might blur together.
Ask for a mix that leans toward 3 to 5:
Mostly ask "explain why" and "what would happen if" questions. Only use multiple choice if I'm stuck.
Step 4: Answer properly, even when you're unsure
Write a real answer, not "idk". A wrong guess followed by the correction sticks better than skipping. Typing your answer also stops you from thinking "yeah, I knew that" when you didn't.
Step 5: Keep a list of misses
Keep a running list of the questions I got wrong or only half right. When I say "review", quiz me again on just those, with the questions reworded.
Rewording matters. If you see the exact same question twice, you may remember the question rather than the idea.
Prompts to copy
- Brain dump first. "Before you ask anything, I'll write everything I remember about this topic. Then tell me what I left out and what I got wrong."
- Exam style. "Write questions in the style of the past paper above. Mark my answers the way an examiner would, and say what would lose marks."
- Teach-back. "Ask me to explain this topic as if you're a classmate who missed the lecture. Interrupt with a follow-up question whenever my explanation skips a step." The Feynman technique guide covers this one in depth.
- Mixed review. "Quiz me on these three chapters in random order, not one chapter at a time." Mixing topics is harder than doing them in blocks, and that's the point.
Mistakes to watch for
Letting it explain first. If you ask "explain photosynthesis" and then "quiz me", you're recalling something you read ten seconds ago. That's short-term memory. Leave a gap, or quiz yourself on yesterday's topic first.
Questions that are too easy. If you're getting everything right, say so: "These are too easy. Ask harder questions that combine ideas from different sections."
Trusting its marking blindly. ChatGPT sometimes marks a wrong answer as right, or makes a confident mistake in its own correction, especially on niche or very technical material. When an answer surprises you, check it against your notes or textbook. Giving it your material in step 1 makes this much less likely.
Doing it once. Recall works best spaced out over days. ChatGPT won't reliably remind you, so put review sessions in your calendar or move your missed questions into a flashcard app.
Where a lesson video fits
Active recall works on material you've already met once. For a topic that's brand new, you need a first explanation before there's anything to recall.
That's where Lesson Video helps. You ask ChatGPT for a short narrated lesson on what you've been discussing, and it plays in the chat with captions and simple animation. Every lesson ends with a question you answer before you see the explanation, so the first recall attempt happens right away. Your answer goes back to ChatGPT, so it knows what to quiz you on next.
A routine that works:
- Watch a short lesson on the new topic.
- Answer its check question.
- The next day, ask ChatGPT to quiz you on it with the prompt from step 2.
- Review your misses a few days later.