Studiare con l’AI: il metodo completo per imparare meglio

Summary: Studying with AI — A Complete Method for Learning Better

The video presents AI as a study companion, not a substitute for studying. Its best role is to help organize work, prompt active thinking, generate practice, and give feedback—while the learner still does the work of understanding and recalling information.

The method is a repeating cycle of activation, deep focus, and repetition. It draws on the presenter’s experience with inattentive ADHD, but is presented as potentially useful to anyone. The presenter also emphasizes that the advice is not a substitute for professional diagnosis or treatment.

Core Ideas

• Studying is active and cyclical. Learning involves making an effort to understand, recall, connect, and revisit material—not just reading or highlighting.
• AI should make the effort more manageable, not remove it. If AI supplies answers too readily, students may feel they have learned without being able to remember or apply the material themselves.
• Motivation matters. Studying only to pass an exam can create anxiety and may not sustain long-term effort. The presenter recommends connecting short-term goals with broader reasons to learn, such as curiosity, practical usefulness, or personal growth.
• The method should match the goal. A basic pass may call for different planning and practice than deep mastery, but the underlying learning principles remain the same.

The Method

1. Activation: Make Starting Easier

The presenter describes the difficulty of beginning as an intention–action gap: knowing what needs to be done does not automatically make it easy to start. AI can act as an external aid for planning and reducing the number of decisions required.

• Write down a reason to begin.
  • Use a short-term goal as immediate motivation—for example, passing an exam or keeping time free later.
  • Over time, try to identify deeper motivations that can sustain continued learning.
• Use an implementation intention.
  • Decide in advance: “If X happens, then I will do Y.”
  • Make X a clear trigger and Y a small, observable action.
  • For example: “If I finish lunch, I’ll go to my desk, open the law PDF to page 3, and read the first paragraph.”
  • The point is to make the decision when you are able to plan, rather than leaving it until the moment you need to act.
• Break large tasks into small actions.
  • “Study anatomy” is too broad to be a practical first step.
  • “Open the PDF and read one paragraph” is easier to start.
  • AI can help break down assignments, courses, or a semester into sessions and first steps.
• Give AI a realistic planning brief.
  • Provide the syllabus, exam date, available study hours, weekly schedule, priorities, and desired level of mastery.
  • Ask for a weekly plan that specifies what to study, how long to spend, what to do next, and a first action that takes less than a minute.
  • The presenter recommends distinguishing topics that need deep mastery from those that only need a solid foundation.
• Try body doubling. Working alongside another person may help with activation and accountability, though the presenter notes it does not work for everyone.
• Create a personal activation ritual. Examples include tidying the desk, using a fidget toy, making coffee, or playing a study playlist. A consistent ritual can become a cue to begin.

2. Deep Focus: Work Actively, Then Allow Mental Breaks

The video describes two broad modes of thinking:

• Focused mode involves directed attention, working through steps, learning new material, and solving problems. It is useful but tiring, and can become unproductive if someone keeps hitting the same obstacle.
• Diffuse mode is relaxed, wandering thought. The presenter associates it with making connections, consolidating learning, and sometimes finding solutions after a break. It cannot be forced, and constant phone or social-media input can turn it into distraction rather than useful downtime.

The modes should be alternated rather than treated as simultaneous. The presenter also discusses ADHD-related hyperfocus and the value of deliberately directing attention toward the task chosen in advance.

Why AI Should Be Guided

The presenter describes a study of nearly 1,000 high-school students in Turkey that compared ordinary study, access to standard ChatGPT, and access to a more constrained AI tutor. As reported in the video:

• During practice, the standard-ChatGPT group solved 48% more problems correctly than the control group; the tutor-style group solved 127% more.
• On a later closed-book test, the standard-ChatGPT group scored 17% lower than the control group, while the tutor-style group performed about the same as the control group.

The presenter’s takeaway is that answer-giving AI can improve practice performance without producing corresponding learning, while an AI tutor that gives hints and guides the learner can avoid that drop. The video cautions that this study concerns mathematics and does not establish that AI has the same effects in every subject.

3. Repetition: Retrieve and Revisit Material Over Time

In the video, “repetition” means returning to material and trying to retrieve it, especially after some forgetting has occurred—not simply rereading it many times.

Retrieval Practice

• After reading, close the book or document and try to explain or recall the material from memory.
• Do not rely on familiarity from rereading as proof that you know something.
• Use AI to:
  • Generate true/false, multiple-choice, short-answer, and application questions from your materials.
  • Give feedback and explain errors.
  • Track recurring mistakes and identify topics to revisit.
• The presenter notes that tracking weaknesses requires deliberate prompting and record-keeping. ChatGPT or Claude memory features may help, but they are not fully reliable or automatic.
• A suggested session prompt is to ask the AI to review previous work, identify and classify mistakes, and recommend what to study next.

Elaborative Interrogation

Elaborative interrogation means asking why and how a fact or process works, and connecting it to what you already know.

Questions might include:

• Why is this true?
• Why does it work this way rather than another way?
• How does it connect to another concept?
• What would happen if the process failed?

Answer in your own words and keep asking follow-up questions. For example, explain why the sympathetic nervous system increases heart rate, then ask why fight-or-flight requires more oxygen and glucose, and continue tracing the explanation.

AI can serve as a Socratic questioner. Its role should be to prompt your reasoning, not to supply a complete explanation immediately.

Explain the Idea to a Beginner

• Explain the topic aloud or in writing while asking AI to pretend it knows nothing.
• Ask it to stop you when you are vague, skip steps, or use unexplained terms.

This helps expose gaps in understanding that can be hidden by fluent but incomplete explanations.

Use Concrete Examples

• Use several specific examples to make an abstract idea easier to visualize and apply.
• First make sure you understand the definition; examples alone cannot replace that foundation.
• Ask AI for examples from different fields, including a few plausible but incorrect examples.
• Decide which examples fit the concept and explain why. This requires applying and distinguishing the principle rather than merely recognizing its wording.

The video illustrates this with negative reinforcement, contrasting examples where an aversive stimulus is removed with cases that instead add a reward.

Spaced Practice, Rest, and Anki

• Distribute study sessions over time instead of concentrating all study into one block and then leaving the topic untouched.
• When returning to material, try to retrieve it before checking the notes. Some difficulty recalling is expected and is part of the practice.
• Rest and sleep matter. The presenter says learning continues between sessions and refers to sleep as supporting the consolidation and reorganization of learning.
• Use AI to generate questions and explain mistakes, but use Anki or another spaced-repetition system to manage when cards should return.
  • Rate how difficult each card was; Anki schedules it sooner or later accordingly.
  • Avoid treating a deck as a passive summary to read through. The useful activity is trying to answer the question before viewing the answer.

Tools Mentioned

• ChatGPT Study Mode: Intended to guide learners with questions, hints, and feedback rather than immediately giving answers. The presenter recommends explicitly asking it to act as a tutor.
• NotebookLM (called “Notukelem” in the auto-generated subtitles): Can create quizzes and flashcards from uploaded sources. The suggested sequence is to load materials, be quizzed, and use the answers to check recall.
• ChatGPT and Claude memory features: May help track goals and recurring errors across sessions, but the presenter advises using explicit prompts and maintaining a written log because the memory is not fully reliable.
• Anki: Schedules flashcard reviews over time based on how well the learner recalls each card.

An Additional Learning Strategy: Explain What You Are Studying

The presenter recommends sharing what you are learning with a classmate, study group, or online audience. The goal need not be to present yourself as an expert; you can document what you understand, where you got stuck, and how you made connections. Preparing to explain a topic encourages deeper questioning, while presenting it requires retrieving it from memory.

If an Exam Is Imminent

If there is very little time, the presenter advises against spending that time building a perfect study system. Instead, prioritize practice questions, correcting mistakes, and reviewing fundamental concepts. For a future exam, start earlier and use the full cycle.

Speakers and Sources Featured or Mentioned

• Ace Spagni — The video’s presenter and the only speaking voice identifiable in the subtitles. He discusses his experience with inattentive ADHD and lifelong learning.
• Unidentified professionals and researchers — The presenter says he interviewed professionals and researchers, but the subtitles do not name them.
• Unidentified study authors — The authors of the reported study comparing standard ChatGPT, a guided AI tutor, and no AI in high-school mathematics are not named in the subtitles.
• Unspecified university studies and research — Referenced generally in the discussion of AI tutors and learning, without individual studies being identified.
• Barbara Oakley — Author of a book the presenter cites in connection with learning and sleep.
• Guang Yang — Named in connection with an experiment about changes to neuronal structures during learning and sleep.
• Cognitive behavioral therapy research for adults with ADHD — Mentioned by the presenter, but no specific researchers or study titles are named.

The tools discussed—ChatGPT Study Mode, Claude, NotebookLM, and Anki—are presented as study tools, not as speakers.