Innovative Cross-Platform AI Education Model: A Case Study of ‘Listening to the World Outside’

Artificial intelligence involves knowledge of big data, voice recognition, machine learning, image recognition, and other technological knowledge. How to teach these concepts to primary school students, how to allow them to experience AI step by step, and what kind of course model to use for AI education are urgent issues to be addressed. Primary school students living in an era of information explosion cannot rely on traditional education models to stand firm in the future era of artificial intelligence.

Currently, many AI education platforms exist at a stage of “separation”. “Separation leads to integration, and integration leads to separation”; AI education must differ from other subjects, including programming education. The project-based learning STEAM+AI cross-platform model allows students to learn autonomous exploratory learning and supports the construction of a tiered ecological system for AI education in primary and secondary schools.

1. Introduction to the Cross-Platform AI Education Model

(1) Theoretical Basis

Artificial intelligence is an emerging technology that simulates human intelligence through perception, reasoning, and behavior. AI not only impacts people’s lives and promotes social development but also advances human thinking. AI education encompasses knowledge from various fields, such as computer science, mathematics, philosophy, physiology, and psychology, making it a highly interdisciplinary emerging discipline. The underlying algorithms of technologies such as machine learning, voice recognition, and image recognition in AI education are very advanced, involving numerous high-tech theoretical knowledge including advanced mathematics. At the beginning of AI education’s popularization, when teachers mentioned AI education to students, they often thought of robots and products far removed from their lives and learning, which undermines the essence of AI education. AI education should not place AI on a pedestal that students cannot reach but instead allow them to find the sources of AI projects from their lives. Therefore, the project-based learning approach, where students learn while playing and play while learning, is very suitable for AI education and aligns with the interdisciplinary education concept of STEAM.

If AI education is taught using traditional methods, dividing chapters and units to help students understand knowledge such as voice recognition and machine learning, it will not only increase the academic burden on teachers and students but also risk that the scientific technologies students use now may be replaced by new technologies in the future strong AI era. AI education must proceed differently from traditional education and even from programming education. The new project-based learning model of AI education enables students to learn autonomous exploratory learning, develop teamwork skills, communication skills, and innovation capabilities, thereby enhancing their intelligence literacy.

In summary, under the guidance of theory and practice, AI education in primary and secondary schools uses a new cross-platform integrated innovative AI education model based on interdisciplinary education concepts. It begins by formulating projects from students’ life and learning contexts, allowing teachers and students to experience AI-related knowledge through various platforms available in the smart education environment. The entire experiential process can take place not only in classrooms but also in STEAM studios, AI family maker spaces, and the surrounding social environment. The series of AI education courses will break down a large project into sub-projects, continuously evolving and iterating the project.

The important model used in the innovative cross-platform integrated AI education model for cultivating students’ innovation capabilities is the “expand-shrink-expand” model. At the beginning of a project, students first find the source of the project from their lives, which is the “expand” phase. When teachers explain basic knowledge, they use a specific tool or platform to solve a project, which is the “shrink” phase. Finally, when students create their projects, they combine real life and learned knowledge to create unique works for their team, possibly using other AI platforms for creation and applying for hardware to extend vertical and horizontal projects, enhancing their autonomous exploratory learning ability, communication skills, and teamwork skills, which is the “expand” phase. The many works produced by students can be displayed, communicated, promoted, and competed, allowing parents to personally experience the entire creative process of their children. The project education under the cross-platform integrated innovative AI education model ensures that every student has an AI product.

(2) Meaning of the Model

Cross-platform refers to the cross-platform for teachers’ teaching and the cross-platform for students’ learning.

Students’ learning: AI has rapidly developed since 2000 alongside deep learning, with the most recognized being CHATGPT, which is just one model. Now major companies like Baidu, Tencent, and Google are researching their own AI education platforms, providing their models on their platforms, which is entirely based on their own hardware and often requires payment. This is a phase of separation; “separation leads to integration, and integration leads to separation”. This book uses cross-platform, where course content is not limited to specific platforms, allowing students to experience AI freely and providing them with ways to experience it for free.

Teachers’ teaching: Utilizing the smart education environment, including public service platforms, Qingdao E platform, and concentrating AI education resources in the teacher’s space for students to access for autonomous exploratory learning. Parents can also see their children’s growth online, which extends into enrichment and advanced courses from this popular course.

In summary, the meaning of the cross-platform model is: teachers, before, during, and after class, respectively identify project needs based on students’ life and learning contexts, leveraging various online learning platforms emerging from the smart education environment to guide students and parents to participate in AI education projects, establishing family maker labs, and promoting exhibitions online and offline showcasing the AI products created by students throughout the process. The AI education environment in the smart environment includes the Qingdao E platform, the central library AI course system, Tencent AI education modules, Puyu education platform, SenseTime platform, public service platforms, school communication, everyone connected space, micro-course connection, teaching assistants, etc.

(3) Stages of the Model

The cross-platform model is basically divided into four stages, as shown in Figure 1.

Figure 1 Stages of the Cross-Platform Model

The first stage is life experience, where only part of the entire educational process is completed in school, and many aspects require students to draw inspiration from life, ultimately applying it to life.

The second stage is AI technology learning, where project-based learning is conducted based on students’ life experiences.

The third stage involves students conducting autonomous exploration and group collaborative learning based on their technical knowledge.

The fourth stage involves group presentations of AI product results, further iterating and refining the works.

2. Introduction to the Cross-Platform Implementation Environment

(1) Hardware Introduction

The curriculum system for AI education is divided into popularization, enrichment, and advanced education. To better allow students to experience AI, the hardware setup can consist of just a computer with a camera and headphones, or a group can add a control board. The use of the control board is not only to experience AI but also to enhance students’ interest in learning and to experience mathematical knowledge. The cross-disciplinary education concept used in this textbook will involve mathematical knowledge, including knowledge of coordinate systems, making it easier for students to understand coordinate knowledge with physical objects.

When conducting project teaching, it is important to note that hardware is used to better implement our project teaching and should not be overly relied upon. The control board, micro:bit, Arduino, and iFLYTEK’s Xiaofei can all be utilized in this textbook. Ultimately, each student’s work is based on different foundations of knowledge.

Throughout the project teaching process, we only use open-source hardware control boards to assist STEAM+AI project teaching, increasing students’ interest in experiences while downplaying the functional role of hardware. This textbook emphasizes the cultivation of original innovation capabilities, including teachers’ original innovation.

(2) Platform Environment

The control board can connect to operations on the Qingdao E platform, and there is a virtual control board available for experimentation on the E platform. In addition to the E platform, MIND+ Baidu services and Puyu openinnolab platform can also connect, with the distinction being that the Qingdao E platform calls the Tencent cloud AI interface, while MIND+ provides the Baidu cloud AI interface. The Baidu cloud interface of MIND+ has service usage limitations, while the Qingdao E platform has VIP restrictions on Tencent AI services. This textbook employs the Qingdao E platform linked to Tencent and the Baidu services of the MIND+ platform, as well as the Puyu openinnolab platform to implement AI courses through cross-platform teaching.

Some existing hardware, platforms, and resources may become obsolete in a few years, but the arrival of AI will not be limited by specific hardware or VIP platforms. The transition from weak AI to strong AI is inevitable. To prevent students from being limited by hardware and platforms, this textbook’s programming implementation or AI experience platform is not restricted; students can use the currently available MIND+ for experience, as well as the Qingdao E platform we are using for experience. While we cannot guarantee that the resources students experience will not become obsolete as they grow up, at this stage, students have utilized these platform resources to experience AI projects and have learned to use platforms for autonomous exploratory learning. Even when they grow up and have new learning environments, students who have experienced the projects in this textbook will still learn, experience, use, and creatively combine with life.

The first half of the textbook projects involve experiencing technologies like voice recognition, image recognition, and the Internet of Things. Once students create intelligent products that make life more convenient and beautiful using these technologies, they may feel that AI technology is impressive and omnipotent. Therefore, in further projects, this textbook is designed to guide students to experience some of the supporting algorithms behind AI. Born in 1956, AI has gone through several peaks and valleys in its development over more than sixty years, influenced by intelligent algorithms, computing speed, storage levels, etc. In recent years, with the rise of big data and enhanced computing power, especially the emergence of new machine learning algorithms, AI has entered an explosive era. Big data + algorithms + computing power is the essence of AI, or large models. Knowledge related to AI should form a tiered curriculum: primary school experience, middle school comprehension, and high school writing.

What will primary school students gain after their experiences? Through teachers’ guidance, students will develop a desire to further understand the underlying data and algorithms, leading to a tiered AI education where they delve deeper into relevant algorithms in middle and high school. Therefore, teachers need to learn about supervised, unsupervised, semi-supervised, and reinforcement learning in machine learning, as well as understand the categories of machine learning algorithms like KNN and decision trees, guiding students to ultimately understand the differences between machine learning methods and human learning methods. Primary school students will no longer view AI through a black box, thinking of it as an unreachable robot, but will be encouraged to actively explore, develop a love for research, and become the new generation of AI leaders.

3. A Case Study: ‘Listening to the World Outside’

The ‘Listening to the World Outside’ project allows students to choose a sound from life, such as their father’s call, the sound of the ocean, or the clanging sounds from a construction site. Students quietly listen to this sound for ten minutes, recording their thoughts and feelings during this time in a STEM practice report. The teacher then explains the basic knowledge of the control board through various platforms and tools available online, teaching students how to convert invisible light and sound into visible shapes displayed on the control board. Subsequently, students experiment with converting the sounds they have heard, during which the teacher guides them to explore knowledge with a questioning spirit, embodying the principle that “to trust books too much is worse than not reading at all.” Through layers of guidance, students’ autonomous exploratory abilities are enhanced. Students hear their father’s love from his call, learn about tolerance from the ocean’s sound, and recognize the importance of learning from the construction site’s noise, thereby enhancing their values and worldview. Students then extend their projects vertically and horizontally, utilizing the control board to apply their projects to life and learning.

The implementation environment of the ‘Listening to the World Outside’ project is open, inclusive, and innovative across platforms. Its project framework is shown in Figure 2, comprising four sub-projects: sub-project one involves students exploring sounds from life, sub-project two investigates the stories behind listening, sub-project three explores the stories behind light, and sub-project four investigates the stories behind the outside world.

Figure 2 Project Framework

The ‘Listening to the World Outside’ project breaks down the boundaries between life and learning, starting from real-life situations and exploring the sources of AI projects, which is the “expand” model. Next, the teacher guides students to utilize the principles of analog-to-digital conversion to implement software components of intelligent light control and sound control lamps in the classroom, as well as to understand the hardware setup of light and sound control lamps through sensors, expansion boards, and control boards, which represents the “shrink” model. Through layers of guidance, students’ autonomous exploratory abilities are enhanced. Based on this, they extend their projects vertically and horizontally, utilizing the control board to apply their projects to life and learning. Students leverage their investigations into life and the information technology knowledge they have learned to unleash their imagination and creativity, completing intelligent systems for their groups, which is the “expand” model. Using the “expand-shrink-expand” model maximizes the preservation of students’ imagination and innovation abilities while guiding them to experience foundational knowledge in class, allowing students to apply their imagination and innovation in project design and creation.

(1) First Stage: Exploration

Students choose a sound from life, such as their father’s call, the ocean’s sound, or the construction site’s clanging noise. They quietly listen to this sound for ten minutes, recording their thoughts and feelings during this time in a STEAM practice report, including reflections on the enhancement of their values and worldview, the practical application of scientific principles, and the improvement of parent-child relationships, as illustrated in Figure 3.

Figure 3 ‘Exploration’ STEAM Practice Report

(2) Second Stage: Students Explore Voice and Image Recognition Technology Based on Their Chosen Projects

Theoretical learning of foundational AI technology knowledge focuses on mathematical concepts like coordinates and analog-to-digital conversion, as well as the drawing of OLED screen line widths, as shown in Figure 4. Cross-platform education encourages both parents and students to actively participate in the activity, achieving home-school co-education, and enhancing students’ self-discipline under the guidance of interest.

The stories behind listening relate to sound, described through the volume and content of the sound. The volume is represented through analog-to-digital conversion, turning the sounds heard into bar graphs, which involves mathematical knowledge and artistic experience. The content of the sound relates to voice recognition, forming control systems, while the stories behind light involve the use of light during the listening process to create control systems. The stories behind the outside world connect to knowledge of image recognition.

Figure 4 OLED Principle

(3) Third Stage: Exploring the Application of Voice and Image Recognition Technology through Vertical and Horizontal Projects

After accumulating foundational knowledge in mathematics and technology, students venture out to listen to the world and engage in vertical and horizontal project creation. From students’ STEAM practice reports, it is evident that through the process of listening to the world, they enhance their values and worldview while also grasping scientific principles, ultimately converting the sounds they hear into images. Subsequently, they continue their autonomous exploratory learning regarding the application of sound in intelligent systems, which constitutes a series of vertical projects. Based on the videos recorded and images captured during their exploration of the world, students engage in intelligent creation, forming a series of horizontal projects.

With a foundational understanding of knowledge and intelligent literacy, the later stages of project-based learning are entirely driven by students’ autonomous exploration, with teachers organizing relevant resources into public service platforms for students to explore based on their interests under teacher guidance.

In both horizontal and vertical projects, teachers no longer need to provide step-by-step instruction on specific programming knowledge. Instead, students engage in autonomous exploration and creation, producing AI products. Teachers consolidate the resources available for autonomous exploration into the teacher’s space, as illustrated in Figure 5.

Figure 5 Teacher’s Space

Students have created intelligent supermarkets, license plate recognition systems, access control systems, smart lamps, intelligent safes, smart trash cans, sound-controlled lamps, light-controlled lamps, and quotation robots, among others. Many projects offer various solutions to problems, allowing me to cultivate students’ emotional intelligence: in learning and life, they will encounter numerous problems and should think of multiple solutions, selecting the most suitable and effective method to actively solve problems. Alongside the education of intelligent knowledge, emotional intelligence, and values, students’ intelligent literacy has significantly improved.

(4) Fourth Stage: Showcasing, Promoting, and Competing with Students’ AI Products

AI education is not a subject where achievements are directly observed through exams; it is more a reflection of students’ comprehensive qualities, such as improvements in other cultural subject scores, logical literacy in problem-solving, and students transferring their love for watching cartoons to autonomous exploratory learning in online learning spaces. Therefore, ‘Listening to the World Outside’ places great emphasis on process evaluation, observing feedback from students, parents, and society through the volume of likes and comments on the platform.

The outputs of this project, including students’ works and interviews with parents, are showcased and evaluated at the International AI Alliance, provincial resource platforms, and provincial maker competitions, as shown in Figures 6 and 7, resulting in numerous awards and honors, and more importantly, enhancing the innovation capabilities and comprehensive qualities of both teachers and students.

Figure 6 Competition Promotion

Figure 7 AI Roadshow Scene

4. Conclusion

The cross-platform integrated innovative AI education model, when implemented again in STEAM+AI, proves that interest is the key to unlocking students’ autonomous exploratory learning, while project-based learning provides wings for AI education, with projects originating from life and ultimately serving life.

AI education is a new discipline, and at the primary school stage, the focus should be on cultivating AI literacy rather than merely imparting AI knowledge. Today’s primary school students are the leaders of the future strong AI era, and their intelligent literacy must be nurtured from a young age, fostering a love for technology, a passion for research, and problem-solving skills. The time for AI education is now; it is not merely programming education or robot education. Under the guidance of a new model, the distinctive features of AI education are highlighted, nurturing the seeds of AI.