Comprehensive Guide to ChatGPT's New Feature "Learning Mode": Learning "How to Think" Instead of Just the Answers—An AI Learning Experience That Presents the Problem-Solving Process

Table of Contents

1. What is "Learning Mode"? —— Background and Aim

2. What's New? Five Core Features

3. Design Philosophy Based on Learning Science

4. How to Use: Activation, Basic Operations, and Tips

5. Specific Learning Flow (Case Study)

6. Usage Scenarios: Schools, Certifications, Language, Workplace Training

7. Differences from Existing Features (Deep Research / Agent / Regular Chat)

8. Frequently Asked Questions (FAQ)

9. Introduction Checklist & Practical Templates

10. Limitations and Risks, Future Roadmap

11. Conclusion
    
    

1. What is "Learning Mode"? —— Background and Aim

OpenAI implemented "Learning Mode" as a new learning experience that emphasizes the problem-solving process, aiming for ChatGPT to function not as a "tool for returning answers" but as a "companion supporting learning." The service will start on July 29, 2025. It will be available on Free, Plus, Pro, and Team plans, and will be rolled out to the Edu plan within a few weeks. The goal is to allow students and learners to focus on **“understanding” rather than “completing”** tasks. OpenAI

2. What's New? Five Core Features

(1) Interactive Prompts
Combining Socratic questioning, hints, and self-reflection prompts, it provides a scaffold for thinking rather than immediately giving answers. OpenAI

(2) Step-by-step (Scaffolding) Answers
Using a section structure that makes connections between key concepts visible, it adjusts to an appropriate amount of context. Even with complex topics, it manages cognitive load, making it easier to deepen understanding through information design. OpenAI

(3) Personalization
It offers optimal difficulty and examples by referencing questions to understand user skills and, if on, memory. OpenAI Help Center

(4) Knowledge Check (Understanding Confirmation)
Through quizzes and feedback on free responses, it supports retention and transfer (application in new situations). You can also incorporate your own materials like PDFs and images. OpenAI Help Center

(5) Flexible On/Off
You can easily switch learning mode on and off during a conversation, allowing you to use it according to your learning goals. OpenAI

3. Design Philosophy Based on Learning Science

The "behavior" of learning mode is supported by custom system instructions designed in collaboration with teachers, scientists, and education experts. The core is the following behavioral principles.

• Encouraging active participation

• Managing cognitive load

  (optimizing the amount of information handled at once)

• Fostering metacognition and reflection

• Stimulating curiosity

• Providing actionable and supportive feedback
  
  
  These are rooted in learning science research and dictate how responses are crafted. OpenAI
  
  

4. How to Use: Activation, Basic Operations, and Tips

Activation

1. Open **"Tools"** in the ChatGPT input field

2. Select **"Study and learn"**

3. Share your goals, level, deadlines, and materials (notes/syllabus/problem images, etc.) to start
   ※You can also start by directly accessing chatgpt.com/studymode. OpenAI Help CenterOpenAI Help Center
   
   
   

Tips

• Initially convey your goals, constraints, and current status

  (what you want to know, by when, and how much you understand)

• Feel free to ask for rephrasing or clarification

• Do not skip knowledge checks; try to explain your reasoning aloud (generative AI is strong in responding to explanation requests) OpenAI Help Center
  
  
  

Turning Off
Uncheck Study and learn from "Tools". OpenAI Help Center

5. Specific Learning Flow (Case Study)

Here, we simulate the behavior of **“step-by-step understanding”** advocated by learning mode using an example from a university foundational course.

1. Goal Setting: "Achieve a passing level on a quiz about 'how to write proofs' in discrete mathematics within a week"

2. Level Assessment: Hearing about prior knowledge, weaknesses, and time allocation

3. Presentation of Learning Plan: Define terms → propositions and proof methods → counterexamples → induction … with step-by-step design

4. Interactive Exercises: Hints/divergence in derivation ("Which definition to use here?" etc.)

5. Knowledge Check: Short questions → free response → immediate feedback

6. Encouraging Metacognition: Articulate where you stumbled and the learning strategy for next time

7. Reflection: Organize the level of achievement and next tasks
   
   
   —This flow aligns with lab examples from OpenAI on game theory, accounting, Bayes' theorem, discrete mathematics, and more. OpenAI
   
   

6. Usage Scenarios: Schools, Certifications, Language, Workplace Training

• Schools (Middle/High School, University): Instead of "outsourcing" homework, it supports the derivation process. Teachers share question design in advance and use the description of the thinking process as an evaluation axis. OpenAI

• Certification Exams: Using "scaffolding" to ensure retention by sequentially understanding definitions → example problems → transformations → limitations of the application range. OpenAI

• Language: Circulating between diagnostic feedback on errors and paraphrasing practice. OpenAI Help Center

• Workplace Training: Incorporate business materials in PDF/images, understanding and applying