Study with AI

Your brain has rules. AI helps you follow them.

Cognitive science has three gold-standard findings about learning: active recall (testing yourself beats re-reading by 50%), spaced repetition (reviewing at increasing intervals moves info into long-term memory), and interleaving (mixing topics beats studying one thing at a time). Most students know none of this. They highlight textbooks and re-read notes — the two least effective study methods according to a 2013 meta-analysis of 10 learning techniques.

AI makes evidence-based studying effortless. Instead of spending 2 hours making flashcards (and never actually reviewing them), you can generate a complete deck in 30 seconds, then spend those 2 hours doing the thing that actually works: testing yourself.

The 30-second flashcard deck that actually works.

Paste your lecture notes, textbook chapter, or slides into AI and use this prompt: "Create 20 flashcards from this material. Each card should test one specific concept. Use the question-and-answer format. Make questions that require understanding, not just memorization. Include 3 cards that connect concepts from different parts of the material."

That last sentence is the secret weapon. Most student-made flashcards test isolated facts. The best flashcards test relationships between ideas — which is exactly what exam questions do. AI is exceptional at finding these connections because it processes the entire material at once.

For tools: Anki is still the king of spaced repetition apps, and you can import AI-generated cards directly. Quizlet has AI features built in now. Or just use a plain text file — the format matters less than the habit of daily review.

Predict the test before you take it.

Here's a prompt that changes everything: "You are a [subject] professor writing a midterm exam. Based on this material [paste notes/syllabus], create a 30-minute practice exam with: 10 multiple choice questions, 3 short answer questions, and 1 essay question. Include an answer key with explanations for why each answer is correct and why the wrong answers are wrong."

This works because AI understands academic testing patterns. It knows that intro biology exams test application of concepts, not just definitions. It knows that history exams ask you to analyze primary sources in context. It knows that calculus exams combine multiple techniques in single problems. Feed it your specific syllabus and lecture notes, and it generates practice tests that are eerily close to the real thing.

Take the practice exam under real conditions: timed, no notes (unless your real exam allows them), no AI. Then grade yourself using the answer key. The questions you get wrong become your study priorities for the next session. This loop — test, identify gaps, study gaps, test again — is the fastest path to mastery.

Explain it to AI like it's five. If you can't, you don't understand it.

Richard Feynman's learning technique: explain a concept in simple language. Where you stumble, you've found your knowledge gap. AI makes this interactive. Open a chat and say: "I'm going to explain [concept] to you. Act as a curious student who asks follow-up questions whenever my explanation is unclear, incomplete, or incorrect. Don't tell me the answer — just ask questions that lead me to the gaps in my understanding."

This is active learning on steroids. You're not passively reading — you're constructing explanations, getting challenged, and rebuilding your understanding in real-time. Students who use this technique consistently score 15-20% higher on conceptual questions compared to re-reading, according to studies on elaborative interrogation.

The beauty of AI here is patience. A study buddy gets bored after 10 minutes. A tutor costs $60/hour. AI will ask you follow-up questions at 3am before your 8am final without complaint. Use that advantage.

Lock it in.

Key concepts to remember.