AI note taking: what helps you learn from handwritten notes
Updated
The short answer
Many AI note takers, such as Otter, record lectures and write transcripts. For notes you write by hand, AI can turn your writing into text, search it, summarize it, and make flashcards and quizzes. The research is strongest for tools that make you recall and practice, spread out over days. Summaries you only reread have weaker evidence.
Two kinds of AI note taker
Many apps sold as AI note takers record a lecture or meeting and write a transcript. Otter captures a class as it happens, then summarizes the key points. Notability has live transcripts on its Pro plan, Goodnotes transcribes audio on its Pro plan and AI Pass, and Notion AI records, transcribes and summarizes meetings. If you want a written record of everything said, one of these is the better pick.
This page is about notes you write yourself, by hand, and what learning research says about each thing AI can do with them. We make alai, a note-taking app for iPad and the web with Ally, an adaptive AI tutor that teaches alongside you, so weigh what we say about alai with that in mind. alai can record audio while you write, and tapping your notes plays that moment, but it doesn't transcribe recordings.
What learning research says about studying
The clearest guide is a 2013 review by John Dunlosky and colleagues, which rated ten common study techniques by how widely their benefits hold up across students, materials and tests:
- High utility: practice testing (quizzing yourself) and distributed practice (spreading study out over time).
- Moderate utility: elaborative interrogation, self-explanation and interleaved practice.
- Low utility: summarization, highlighting, the keyword mnemonic, imagery for text and rereading.
In a classic experiment by Roediger and Karpicke (2006), college students learned a short passage by reading it four times, reading it three times and recalling it once, or reading it once and recalling it three times. On a test five minutes later, the readers did best. A week later, the students who had recalled it three times remembered 61% of its ideas and the readers 40%, though the readers had felt surer they would remember it. For spacing, a 2006 synthesis by Cepeda and colleagues pooled 839 assessments from 317 experiments and found that the best gap between study sessions grows the longer you need to remember.
What AI can do with handwritten notes
| AI feature | What it does | Evidence for learning |
|---|---|---|
| Handwriting to text or LaTeX | Makes a typed copy | A convenience |
| Search inside handwriting | Finds the page you need | Helps you review |
| Summaries | A note's main points | Weak if you only reread them |
| Flashcards and quizzes | Questions from your notes | Strong: practice testing |
| Spaced review | Cards come back when due | Strong: distributed practice |
| A tutor that follows your work | Responds to the step you wrote | Strong for step-by-step tutors |
Handwriting to text, LaTeX and search
Typed text or LaTeX is handy for writing up a lab report, and search saves flipping through a semester of pages. Neither is a study technique, but both get you back to your notes, which matters: a 2006 meta-analysis of 33 studies found that taking and reviewing notes had substantial effects on learning compared with doing neither.
Summaries
An AI summary shows what a long note covers. As a way to study, it has weak support. Dunlosky's team rated summarization low because its benefits showed up only for some students on some tests, and rereading low because it doesn't consistently raise scores. Reading a summary someone else wrote is closer to rereading. Handing the thinking to AI can also make the work too easy: in a 2024 experiment with 91 university students, those who researched a topic with ChatGPT had a lower cognitive load than those who used Google, and backed their recommendations with weaker reasoning.
Flashcards and quizzes
These are practice testing: they make you pull the answer out of memory before you see it. One caution applies to cards made for you. In six experiments by Steven Pan and colleagues, students who wrote their own flashcards remembered more two days later than students given ready-made cards (d = 0.45), and in most experiments they did better on questions that asked them to apply the material (d = 0.29). That study didn't test AI, but cards an AI writes for you are ready-made too. Start from the AI's cards, rewrite the ones you miss in your own words, and add a few of your own.
Spaced review
A schedule that brings each card back when it's due does the spacing for you, and Cepeda's synthesis suggests longer gaps for things you need to remember longer, such as a final exam.
A tutor that follows your work
In Kurt VanLehn's 2011 review, computer tutors that mostly respond to each step of a problem had an effect size of 0.76 compared with no tutoring, close to human tutors at 0.79. With AI chat, design matters: in a 2025 PNAS field experiment with nearly 1,000 high school students, those who practiced math with a ChatGPT-style chat scored 17% lower on a later exam than students who never had AI, while a version with prompts designed to protect learning largely avoided that loss. The guide to homework help that gives hints instead of answers covers that study.
Does writing by hand help?
Maybe a little. In 2014, Pam Mueller and Daniel Oppenheimer found in three studies that students who typed lecture notes on laptops did worse on conceptual questions than students who wrote by hand, and linked that to laptop notes copying the lecture word for word. Later work was less clear. Morehead, Dunlosky and Rawson (2019) found no consistent difference and called it premature to say which method is better, and a 2021 direct replication by Urry and colleagues with 142 students found no quiz advantage for longhand. A 2024 meta-analysis of 24 studies by Flanigan and colleagues found a small advantage for handwritten notes on college students' achievement (g = 0.25), while typing produced far more notes (g = 0.92). Either way, testing yourself and spacing your review still apply.
What to look for in an AI note-taking app
- It makes you answer. Flashcards, quizzes and a tutor that asks questions have the most evidence behind them.
- It brings material back over days. Spaced review beats one long sitting.
- It shows where each card came from. AI can get things wrong, so check cards against the page they came from.
- It handles your handwriting. Try it on your real notes, and on your math if you take STEM classes.
- It's clear about your data. Check where recordings are kept and whether your notes are used to train AI models.
Which kind of tool fits you
- You want a transcript of every lecture. A transcription tool such as Otter fits best. Otter records from a phone, a computer or a Chrome extension, transcribes live, and has a free Basic plan. Check your school's rules about recording a class.
- You take typed notes. Notion AI works inside typed pages, with drafting, meeting notes and search across your workspace. Notion's Free plan includes a trial of it.
- You write by hand. Goodnotes' AI can search handwriting, summarize, turn handwriting into text, check handwritten math and make quizzes from notes. Notability Learn makes summaries, quizzes, flashcards and fill-in-the-blanks from notes, PDFs and recordings on Notability's Plus and Pro plans. alai is below.
How alai handles the parts that help
alai's note tools are on every plan, including Free: pens and a highlighter, Cornell and other papers, PDFs you can write on, and tape, which covers part of the page until you tap to peek. Tape makes a simple flashcard of your own, and it fits the recite step of Cornell notes. Free records 20 minutes of audio per note, kept on your device, and gives you 30 minutes, once, to try Ally.
Pro, at $14.99 a month or $119.99 a year, adds the AI note tools and Ally:
- Flashcards from any note, in any class. You flip a card and mark it Know it or Still learning, and flashcard review brings cards back when they're due.
- Quizzes you answer by hand or by typing. Ally marks each answer right, close or not yet, gives a short hint when you miss, and shows the answer only when you ask.
- Page links. Each summary point, card and quiz question links to the page it came from.
- Summaries, handwriting to text or LaTeX, and search inside handwriting, plus study games.
- Ally, 10 hours a month per learner. In Homework help, Ally checks each step about two seconds after your pen lifts, stays quiet while the work is right, and at the first step that doesn't follow asks a short question, such as “Nice job getting 5x; does the 5 in 5(x + 2) reach the 2 too?” Ally remembers what was tricky last time and can make practice problems from your past slips. You can see, edit or delete what Ally remembers.
Ally tutors math from algebra through calculus, physics, chemistry and Spanish, and the study tools work on notes from any class. Ally can misread messy handwriting or judge a step wrongly, and alai hasn't published any measurements of how much students learn with the app. alai doesn't train AI models on your notes. The web app runs in a browser such as Chrome or Edge, so it works on a Chromebook or a Windows laptop, and the iPad app needs iPadOS 26 or later. Pricing lists every feature by plan.
Questions
What is the best AI note taker for students?
It depends on how you take notes. If you want a transcript of every lecture, a transcription tool such as Otter fits. If you write your notes by hand, look for an app that turns them into flashcards and quizzes and brings cards back over days, since practice testing and spacing have the strongest evidence.
Can AI turn my handwritten notes into flashcards?
Yes. Notability's Learn tool makes flashcards from notes on its Plus and Pro plans, and alai Pro makes flashcards from any note and brings them back when they're due. In one set of experiments, students who wrote their own cards remembered more, so rewrite the AI's cards you keep missing in your own words.
Does alai transcribe lectures?
No. alai can record audio while you write, and tapping your notes plays what was said at that moment. The recording is kept on your device, and alai doesn't turn it into text or make notes from it.
Are AI summaries good for studying?
They're useful for seeing what a note covers. As a study method, reading a summary is close to rereading, which has weak evidence, so test yourself on the material afterward with flashcards or a quiz.
Is it better to take notes by hand or on a laptop?
The evidence is mixed. A well-known 2014 study favored handwriting, later replications found little or no difference, and a 2024 meta-analysis found a small advantage for handwritten notes. Testing yourself and spacing your review help whichever way you take notes.
Sources
- Dunlosky et al., Improving students' learning with effective learning techniques: Promising directions from cognitive and educational psychologyPsychological Science in the Public Interest, 14(1) · January 2013
- Roediger and Karpicke, Test-enhanced learning: Taking memory tests improves long-term retentionPsychological Science, 17(3) · March 2006
- Cepeda et al., Distributed practice in verbal recall tasks: A review and quantitative synthesisPsychological Bulletin, 132(3) · May 2006
- Kobayashi, Combined effects of note-taking/-reviewing on learning and the enhancement through interventions: A meta-analytic reviewEducational Psychology, 26(3) · June 2006
- Stadler, Bannert and Sailer, Cognitive ease at a cost: LLMs reduce mental effort but compromise depth in student scientific inquiryComputers in Human Behavior, 160 · November 2024
- Pan et al., User-generated digital flashcards yield better learning than premade flashcardsJournal of Applied Research in Memory and Cognition, 12(4) · December 2023
- VanLehn, The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systemsEducational Psychologist, 46(4) · October 2011
- Bastani et al., Generative AI without guardrails can harm learning: Evidence from high school mathematicsProceedings of the National Academy of Sciences (PNAS) · June 25, 2025
- Mueller and Oppenheimer, The pen is mightier than the keyboard: Advantages of longhand over laptop note takingPsychological Science, 25(6) · June 2014
- Morehead, Dunlosky and Rawson, How much mightier is the pen than the keyboard for note-taking? A replication and extension of Mueller and Oppenheimer (2014)Educational Psychology Review, 31 · February 4, 2019
- Urry et al., Don't ditch the laptop just yet: A direct replication of Mueller and Oppenheimer's (2014) Study 1 plus mini meta-analyses across similar studiesPsychological Science, 32(3) · March 2021
- Flanigan et al., Typed versus handwritten lecture notes and college student achievement: A meta-analysisEducational Psychology Review, 36 · July 12, 2024
- Otter for educationOtter.ai · accessed November 5, 2026
- Otter.ai home page and plansOtter.ai · accessed November 5, 2026
- Notability LearnNotability Help Center (Ginger Labs) · accessed November 5, 2026
- Goodnotes AIGoodnotes · accessed November 5, 2026
- Notion AINotion · accessed November 5, 2026