China Insights Consultancy: AI Education Report Summary

China Insights Consultancy Report

About CIC 灼识咨询

  • A well-known consulting firm offering services such as IPO industry consulting, commercial due diligence, and strategic consulting.
  • Expertise in consumer goods, internet, big data, high technology, energy, AI, financial services, healthcare, education and more.
  • Utilizes primary research (expert interviews) and secondary research (publicly available data from sources like the National Bureau of Statistics of China).
  • Employs internal data analysis models to ensure accuracy, using data up to the report's publication date.

Contact Information

  • For detailed information, media/marketing cooperation, or to join industry groups, scan the QR code or email marketing@cninsights.com.

Key Insights: AI and Education

  • Core Idea: Large models are reshaping the future of education, with intelligent adaptive education leading the way in AI+Education.
  • Market Size (China Adaptive Learning Machines):
    • 2023: Approximately ≈40 billion RMB\approx 40 \text{ billion RMB}
    • 2028E (Estimated): > 200 \text{ billion RMB}
  • China Education Intelligent Hardware Market Size:
    • 2023: 430 billion RMB
    • 2028E (Estimated): 790 billion RMB
  • Market Share: Adaptive learning machines are expected to exceed 60% market share.
  • Global Impact: Adaptive education is becoming a key force in the globalization of AI education, with over 300 million users worldwide.
  • Leading the Trend: Squirrel AI has launched the world's first adaptive education large model, pioneering the commercialization of intelligent teachers.

Report Chapters

  1. Overview: Global Education Innovation Wave
  2. In-Depth Analysis: Learning Machines Dominate the Hardware Market
  3. Insight: Adaptive Learning Machines as Market Darlings
  4. Practice: First Adaptive Education Large Model Implementation

Chapter 1: Large Model Applications Lead the Way

  • Learning Efficiency Enhancement: Traditional Education + Adaptive Technology + Large Models.
  • Ideal Application: The education sector is ideal for large model technology due to data resources, personalized learning demands and social value.
  • Comprehensive Empowerment: Large models are empowering the entire education chain, leading to intelligent and personalized education.
  • Personalized Learning: Adaptive education, with AI and big data, creates personalized learning blueprints.
  • Key Enabler: Large models drive adaptive education, enabling accurate demand analysis and personalized learning paths.
  • Global Trend: AI+Education is a global trend driven by capital and policy support.
  • Global Adoption: Adaptive education has reached nearly 300 million people globally.

Education Industry Development

  • Progression: From traditional teaching to intelligent adaptive learning.
  • Stages:
    • Traditional teaching (before 2000):
      • Large classes, small classes, 1v1 teaching
    • Efficiency Improvement (2000-2012):
      • Two-teacher classes, online live broadcasts, intelligent grading
    • Decision-Making Assistance (2012-2016):
      • Adaptive question banks, emotion recognition, AI teaching assistants
    • Intelligent Adaptive Teaching (2016-present):
      • Adaptive learning systems, AI smart teachers, adaptive education models, multimodal adaptive education models, source tracing technology
  • Industry Background: Internet era, increased education investment, emergence of education companies like New Oriental.

Global AI+Education Development

  • Timeline:
    • 1954: Skinner's behaviorist learning theory.
    • 1956: Dartmouth Conference, the birth of AI.
    • 1960: First computer-aided teaching system, PLATO.
    • 1965: DENDRAL, first expert system.
    • 1970s-1990s: Development of intelligent tutoring systems.
    • 2006-2022: Deep learning advancements (Hinton, CNN, AlphaGo, ChatGPT).
    • 2023-2024: Release of domestic education models and adaptive education models.
  • Trends: Scalability, fairness, personalization, standardization.

General Large Model Development

  • Evolution: From basic understanding to logical reasoning.
  • Stages:
    1. Basic Understanding: Generates low-quality content.
    2. Understanding Human Preferences: Generates high-quality content.
    3. Logical Reasoning: Generates near-human level content.
  • Examples: LeNet-5, GAN, Google Bert, GPT series, Huawei Pangu, Baidu Wenxin, Kimi, Sora.

Large Model Application Scenarios

  • Education Focus: Personalized learning, AI teachers, intelligent grading, education data mining, intelligent Q&A.
  • Other Scenarios: Digital business halls, report generation, product recommendation, intelligent investment consulting, intelligent marketing, intelligent risk control.

Education as an Ideal Application for Large Models

  • Key Advantages: Rich data, personalized learning needs, willingness to pay, high social value.
  • Data Resources: Learning data, teaching methods, educational content.
  • Personalization: Analysis of learning behavior, speed, style, and performance.
  • Social Value: Promotes educational equity, innovation, quality, efficiency and knowledge.

Understanding Adaptive Education

  • Definition: Education tailored to individual characteristics using AI and user data.
  • Levels by Intelligence: Traditional education, internet education, intelligent tools, partial adaptive education, advanced adaptive education, fully adaptive education.
  • Teacher Roles: From teaching to mentoring depending on the level of AI integration.

Role of Large Models in Adaptive Education

  • Benefits: Semantic understanding, knowledge representation, personalized analysis, guided learning, full subject knowledge integration.
  • Impact: Accelerates intelligent education and personalized teaching.

Global Investment in AI+Education

  • Investment Trends: Increased investment in AIGC+Education projects globally.
  • Investment by Country: Primarily in the US, India, UK, China, Finland, Italy, and Spain.
  • Investment by Stage: Seed/angel rounds, A rounds, B-C rounds.
  • Investment by Sector: K-12, vocational training, comprehensive applications, language learning, higher education, STEM education and early childhood education.

Global Policies and Initiatives in AI+Education

  • UNESCO: Guidelines for Generative AI in Education and Research (2023).
  • European Union: AI Act draft (2023) classifying AI systems in education as high risk.
  • Australia: Generative AI in Australian Schools framework (2023).
  • United Kingdom: Generative AI in education (2023).
  • Japan: Provisional guidelines for AI use in education (2023).
  • United States: AI and the Future of Teaching and Learning (2023).

Global Adaptive Learning Platforms

*Note: User numbers are estimates.

  • Market Penetration: Adaptive learning has been adopted by approximately 300 million users.
  • Platforms Included: DreamBox Learning, Mathspace, Fast ForWord, Lexia Learning, ABCMouse, Adventure Academy, Homeschool+.

海外头部智适应学习平台概览 (2/2)

  • DreamBox Learning:
    • Provides digital math and reading courses with an adaptive learning system.
    • Studies show improved math skills with DreamBox Learning.
  • Mathspace:
    • Offers personalized guidance and assessment using AI and machine learning.
    • Optimizes the learning process.
  • Fast ForWord:
    • Increases skilled and advanced readers and reduces struggling readers.
    • Improves STAAR math scores.
  • Age of Learning (ABCMouse, Adventure Academy, Homeschool+):
    • Commits to integrating education best practices, innovative technologies, and creativity into educational programs.
    • Offers effectiveness that’s proven through rigorous efficacy research.

Chapter 2: Learning Machines Lead Intelligent Hardware

  • Definition: Education Intelligent Hardware = Traditional Electronic Devices + Quality Education Content + Cutting-Edge Digital Technology
  • Market Growth: The market reached 430 billion RMB in 2023 and is expected to exceed 790 billion RMB by 2028, with an annual compound growth rate exceeding 13%.
  • Hardware Leader: Learning machines stand out for their functions, application and value and stand as core driver of industry growth.
  • Market Competition: Traditional hardware manufacturers and cutting-edge technology companies such as Squirrel AI are leading innovation with adaptive education and large model technology.

Education Equipment Industry Overview

  • Stages: Beginning, exploration, growth and development phases.
  • Key milestones: The emergence of Xiaobawang learning machines, launch of Wenquxing and Hanwang products, introduction of Du Shulang point-and-click machines, NetEase dictionary pens, Squirrel AI 5G + AI learning machines, Baidu “Xiaodu Smart Learning Tablet”.

Education Intelligent Hardware Classification

  • Products Included: Learning machines, early education robots, dictionary pens, incorrect question printers, smart desk lamps, smart watches, point-and-read pens, and word cards.
  • Learning Machine Value-Add: Learning inside and outside the curriculum, breaking down knowledge points, dynamic Q&A, and providing solutions, including mistake anaylsis.

China Education Intelligent Hardware Market Size

  • Market Value: 2019-2028E
    • 2019: 252 billion RMB
    • 2023: 430 billion RMB
    • 2028E (estimated); 790 billion RMB
  • CAGR: Growth is attributed to advances in technology which improve the performance and interactivity of learning resources.

Role of Education Intelligent Hardware

  • Key Role: Providing personalization is the key role in the digital transformation necessary to scale education.

Factors Driving Growth in Education Intelligent Hardware

  • Beneficial to Stakeholders: Education intelligent hardware acts as a helper to students, parents and schools.
  • Catalyst for Digital Change: Digital education, combined with a personalized model is at the forefront of these transformation.
  • “Double Reduction” Policy: The "double reduction" policy has transferred high-quality education needs leading to demand for intelligent education hardware.

Key Players

  • Included: Squirrel AI, NetEase Youdao, Baidu, iFlytek, Homework Help, Xueersi, Youxuepai, Dubulang, Yuan Exercise, BBK (Bu Bu Gao).

Key Players in Intelligent Devices

  • Evolution: From technology development to industry layout.
  • Dominance: Squirrel AI distinguishes itself by being the education technology company to conduct human-machine battle experiments, achieving numerous victories thanks to its adaptive education system.

Overview of Empirical Global Results

  • Human-Machine Battle: Globally, human-machine battles prove that adaptive education improves understanding.
  • Study:
    • Squirrel AI’s average score improvement was more than 5×5\times that of a human teacher in the same period, and the increase in knowledge mastery was more than that 3×3\times in over 100 human-machine battles
    • City Springs Public Middle School in Miami reported that after using Waggle, the proportion of students passing the Northwest Assessment Association’s teaching effectiveness test in reading subjects increased by 25%.

Chapter 3: Adaptive Learning Machines Emerge

  • Definition: Adaptive Learning = Traditional Learning Machine + Intelligent System + Large Model.
  • Market Strength: In line with current AI technology trends. With features such as precise diagnosis, personalized recommendation, and source tracing, adaptive learning machines are becoming increasingly popular among students.

Increased Market Share

  • Key Trend: The learning machine market is in a transitional period, adaptive solutions that incorporate features such as personalized AI tutors and precise diagnostics are expected to have a market share greater than 60% by 2028.

Education Intelligent Hardware Product Needs

  • Wide Applicability: Learning machines are applicable to a wide age group, starting from age 3 to 25 years old, making the value very accessible and are likely to see continuous advancement.

China Learning Machine Market Details

  • Benefits: A high growth rate is attributed to artificial intelligence which enables adaptive learning, personalized recommendations and efficient learning.

Evolution Details

  • Characteristics: From intelligent recognition to adaptive machine.

Adaptive Learning Machine Technical Infrastructure

  • Layers of Analysis: Data Processing from user feedback enables a better understanding of requirements and therefore improves teaching ability.

Technical Breakdown

  • Data Layer: The Adaptive Learning machine incorporates analysis of Students History and Full Subject Nano-Level Knowledge.
  • Model Layer: Education Database, Education Intelligent Body, Enhanced Generation etc.
  • Application Layer: AI Root Cause Tracing, Misunderstanding Analysis , Personalized recommendations and Feedback.

Intelligence Creation

  • Key Creation: Improves personalized intelligent capabilities and problem solving.

C-Side Value Additions

  • Includes Comprehensive Learning, strong interactivity and Personalized customization.

Market Size

  • Analysis for intelligent education solutions: intelligent solutions based in current technology, are taking up an increasing market share, and seeing rapid growth.

Chapter 4: Education Reform

  • AI Integration
    • AI is changing the field, integration with large language models improves a variety of dimensions in education that were not previously possible.

Matrix

  • An AI Learning System, powered by a strong data set, is integral to success.

Model Access Approaches

  • Contrast: A multi-faceted approach, some models employ APIs, while others import from 3rd party sources.

Squirrel AI

  • Details: A market leader with a distinct brand and many smart machine clients.

Squirrel AI Adaptive Education General Model

  • Layers: A framework that allows machine learning and user integration.

Adaptive System Capabilities of Squirrel AI Based Large Models

  • Breakdown: The capacity for dynamic questioning, systematic capabilities and personalized assistance enables Squirrel AI to deliver individualized value.

Squirrel AI innovation And Collaboration Power

  • Dominance: A proven success in field testing, Squirrel AI has a strong innovation background with many authorized patents as well being featured in the Harvard Business Review's case study library, and earning the Ram Charan Innovation Practice Award.

Core Takeaways:

  • Adaptive educational systems based in large language models, have a proven success record as well as strong backing from academia and government entities.