Develop AI Teaching and Learning Resources for Compulsory Education in China

Jiachen Song, Jinglei Yu, Li Yan, Linan Zhang, Bei Liu, Yujin Zhang, Yu LuAAAI 2023 · DOI

TL;DR

This paper addresses the critical need for AI educational resources in China's compulsory education system, which has recently mandated AI instruction. The authors developed a comprehensive AI knowledge structure and a suite of free, open-source teaching and learning materials, including lesson plans, interactive activities, and coding projects, to support teachers and students from grades 1-9. Their work has been implemented in teacher training programs and pilot school courses, demonstrating significant improvements in student AI knowledge and computational thinking.

What Problem Does This Paper Address?

China has recently made AI a mandatory subject within its compulsory education curriculum for students in grades 1-9. This presents a significant challenge because many teachers and schools are ill-prepared to teach AI. The primary obstacles include a severe lack of adequate teaching and learning resources, which are time-consuming and expensive for educators to develop from scratch. Furthermore, many in-service teachers lack the necessary AI-related content knowledge, technological skills, and pedagogical expertise to effectively teach the subject. This gap leaves millions of students without the foundational AI education they are now entitled to.

Methods

The authors developed a structured approach to address the resource gap. First, they created an AI knowledge structure that aligns with China's national curriculum for information science and technology. This structure is organized into six modules and is compatible with the "five big ideas" of AI for K-12 education (perception, representation and reasoning, learning, natural interaction, and societal impact).

Based on this knowledge structure, they designed a comprehensive set of 27 teaching and learning resources. These resources are built around project-based learning and aim to foster key literacies such as computational thinking and information society responsibility. The resources are designed to be free and accessible online, reducing the need for schools to invest in specialized software or hardware. They include a variety of formats such as:

  • Lesson Plans: Structured for backward design, focusing on real-world problems and student inquiry.
  • Slides: Covering introductions, new concepts, and hands-on practice.
  • Questionnaires: To guide student inquiry and awareness of AI's societal impact.
  • Warm-up Games: To increase student motivation and engagement.
  • Demonstration Videos: To explain complex concepts or tasks.
  • Flow Charts: To help students visualize and plan algorithms before coding.
  • Record Tables: To assist students in organizing data from experiments or activities.
  • Quizzes: To assess understanding and reinforce learning.
  • Rubrics: For evaluating project outcomes.
  • Instruction Manuals: For self-learning and hands-on projects.
  • Codes: Including both block-based programming (e.g., for Micro:bit) and Python-based projects using platforms like Tencent Coding, which supports deep learning frameworks.

Datasets

Not discussed in the paper. The paper focuses on the development of educational resources rather than training AI models on specific datasets.

Key Insights and Ideas

  • Bridging the Gap with Open Resources: The most significant insight is the successful creation and deployment of free, open-source AI educational resources tailored to a national curriculum. This approach democratizes AI education by removing financial and technical barriers for schools and teachers.
  • Project-Based Learning as a Core Pedagogy: The emphasis on project-based learning is crucial. It allows students to engage with AI concepts through practical application, fostering deeper understanding and problem-solving skills, rather than rote memorization.
  • Teacher Training is Paramount: The paper highlights that simply providing resources isn't enough. A robust teacher training program is essential to equip educators with the confidence and skills to effectively integrate AI into their classrooms. The authors' program addresses specific teacher-identified challenges.
  • Alignment with National Curriculum is Key: The AI knowledge structure's direct mapping to China's national curriculum ensures relevance and facilitates adoption by educational institutions.
  • Iterative Development and Verification: The process of implementing exemplary courses in schools and collecting feedback (e.g., student satisfaction, knowledge gains) demonstrates a commitment to refining and validating the educational materials.

Key Results

  • Comprehensive Resource Suite: The developed AI knowledge structure includes 90 core concepts, 63 learning indicators, and 27 distinct teaching and learning resources.
  • Teacher Training Impact: Over 900 teachers have been trained through their program, with participating teachers reporting high satisfaction (over 95% for various aspects like lectures, topics, and resources).
  • Classroom Implementation Success: In a pilot study with 20 fourth-grade students over 8 weeks, participants showed significant improvements:
  • AI Knowledge: A statistically significant increase (t=-2.268, p=0.035) was observed.
  • Computational Thinking: Improved problem-solving skills were noted (t=-2.153, p=0.044).
  • Student Engagement: Students expressed strong interests in AI (mean=4.074) and high satisfaction with the teaching methods (mean=3.74).
  • Resource Accessibility: The online platform (https://coding.qq.com/ai/) provides free access, saving schools investment in software and hardware.

Limitations and Future Work

  • Scalability of Teacher Training: While 900 teachers have been trained, the goal is to train at least 3,000 AI teachers in three years. Scaling this training effectively, especially post-COVID, remains a focus.
  • Resource Iteration: The authors plan to develop more courses, verify their effectiveness, and iterate on the AI knowledge structure, drawing inspiration from international K-12 AI curricula.
  • Encouraging Design-Based Research: They aim to encourage more teachers to participate in design-based research to accelerate the development of customized AI courses.
  • Broader Implementation: While a pilot study was successful, broader implementation and evaluation across diverse school settings in China will be crucial.

Summary updated

BibTeX
@inproceedings{song2023develop,
  author = {Jiachen Song and Jinglei Yu and Li Yan and Linan Zhang and Bei Liu and Yujin Zhang and Yu Lu},
  title = {{Develop AI Teaching and Learning Resources for Compulsory Education in China}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2023},
  doi = {10.1609/aaai.v37i13.26904},
}
Figure