Currently, AI education in primary and secondary schools has shifted from “whether to implement it” to “how to implement it effectively.” However, in reality, some schools still exhibit a “one-size-fits-all” approach in their curriculum design. Some teachers directly copy university textbooks, filling the classroom with abstract algorithm formulas; some schools are eager to introduce high-tech equipment but overlook students’ actual understanding capabilities; and some even simplify AI education to mere “programming classes” or “robotics classes,” drifting away from the core competency cultivation goals. The root of these issues often lies in the neglect of cognitive development principles. In fact, AI education, as an emerging discipline that integrates technology and thinking, must be based on a deep understanding of students’ cognitive characteristics. Only by following the growth logic from the concrete to the abstract, from the perceptual to the rational, and from interest to inquiry can AI education truly take root and become fertile ground for stimulating students’ innovative potential.
1. Lower Grades of Primary School: Igniting Interest with “Tangible Intelligence”
Students in the lower grades of primary school have distinct cognitive characteristics; their thinking is primarily concrete and visual, with short attention spans, showing more interest in things that are “visible, tangible, and interactive”; understanding abstract concepts relies on life experiences, and their inquiries about “why” often stem from intuitive feelings. The key to AI education at this stage is not to teach the definition of “what artificial intelligence is,” but to allow students to perceive the existence of “intelligence” through “tangible intelligence,” igniting their intrinsic motivation to learn more.
For example, a first-grade science class once designed a theme activity called “A Day with an Intelligent Partner.” The teacher did not rush to explain sensors or algorithms but instead brought three “mystery boxes”: the first box contained a talking smart speaker, which played music when students said, “Play a nursery rhyme”; the second box was a smart desk lamp with infrared sensing, which automatically lit up when someone approached and slowly dimmed when they left; the third box was a vacuum robot, which the teacher had programmed in advance to “patrol” the classroom, automatically turning when it encountered a table corner. Students gathered around the boxes, operating them while exclaiming, “Wow, it can ‘hear’ me talk!” “How does it know someone is coming?” “Why doesn’t it bump into the table?” The teacher then guided students to observe “intelligent friends” in their lives—like smart TVs at home, automatic door access in their community, and self-checkout machines in supermarkets—making them realize that “intelligence” is not an unreachable technology but a “helpful assistant” hidden in everyday details. After class, students eagerly drew their “ideal intelligent backpack”: some illustrated backpacks that could automatically organize books, others depicted backpacks that reminded them to bring homework, and some even imagined backpacks that could measure temperature. These imaginative ideas full of childlike wonder are the starting point of AI education—not teaching knowledge but planting the seeds of curiosity.
This stage should avoid two misconceptions. One is the excessive pursuit of “technical content,” using complex devices or terminology that intimidate students; the other is the forced indoctrination of the concept of “artificial intelligence,” leading students to lose interest due to abstraction. Teachers should focus on “perception” and “experience,” using gamified and life-related activities to help students establish a preliminary understanding of “intelligence” through play, accumulating perceptual experiences for future learning.
2. Upper Grades of Primary School: Understanding Principles through “Problem Solving”
As students enter the upper grades of primary school, their cognitive abilities significantly improve. They begin to transition from concrete thinking to abstract logical thinking, can understand simple cause-and-effect relationships, and develop a strong curiosity about “how to do it” and “why”; at the same time, their hands-on skills and collaborative awareness increase, and they desire to validate their ideas through practice. AI education at this stage should move beyond the level of “perception” to “understanding”—by solving real problems, students can experience the core principles of AI in practice and establish a mindset of “technology serving needs.”
For example, a fifth-grade project called “Intelligent Librarian” was designed based on this logic. The project stemmed from a real issue faced by students, where the class library corner often led to conflicts due to disorganized classifications and difficulty finding books. The teacher guided students to think, “If there were an ‘intelligent librarian’ that could help us quickly find books, wouldn’t that be convenient?” Students immediately began discussing and ultimately defined the requirements: this “librarian” needed to recognize book categories (such as fairy tales, science books), remember the location of each book, and be able to “speak” to guide students when they inquired.
Next, the teacher advanced the course in three steps. First, combining knowledge from science class about “classification and organization,” students were guided to establish book classification standards (such as by color labels or barcodes); then, integrating knowledge from math class about “coordinate positioning,” students learned to mark the library corner’s location on a grid map, describing positions using “rows and columns”; finally, combining knowledge from information technology class about “sensors and programming,” students used Arduino development boards and infrared sensors to attach electronic tags to each book. When students said the book title, the sensor would receive the signal and trigger a voice prompt indicating the book’s location. Throughout the process, students needed to continuously debug the program—for example, when two books were too close together, the sensor could easily misjudge, so students tried optimizing it using the “multiple detections to take the average value” method; when the voice prompt was not clear enough, they adjusted the microphone’s position and volume parameters. By the end of the project, students not only completed a functioning “intelligent librarian” model but also deeply understood the basic logic of AI: “data collection – algorithm processing – result output.” One student wrote in their diary, “I realized AI is not magic, but a way to solve big problems with many small steps!”
The key at this stage is problem-driven and interdisciplinary integration. Teachers need to extract real problems from students’ lives, break them down into manageable sub-tasks, and guide students to comprehensively apply knowledge from math, science, information technology, and other subjects to naturally understand AI principles while solving problems. At the same time, students should be allowed to “trial and error”—for instance, program failures or inaccurate models; these “errors” are precisely the best opportunities to understand the limitations of technology.
3. Middle School Stage: Cultivating Innovation through “Open Inquiry”
Middle school students’ cognitive development enters the stage of formal operations. Their abstract thinking abilities significantly enhance, allowing them to understand complex logical relationships and possess the ability for hypothesis deduction; at the same time, they begin to focus on social issues, eager to use their knowledge to solve real problems, with a stronger demand for “innovation” and “critique.” AI education at this stage should transcend the level of “understanding principles” to “application and innovation”—guiding students to design AI solutions based on real needs, cultivating critical thinking and social responsibility through inquiry.
For example, an eighth-grade project called “AI for the Elderly” is a typical practice of this idea. The project originated from a community survey where students discovered that 35% of the community’s elderly population was over 60 years old, and some elderly individuals faced difficulties using smart devices (such as mobile payments and health monitoring) due to declining vision and hearing. Students then proposed, “Can we design an AI assistant suitable for the elderly to help them solve practical problems in their lives?”
Under the teacher’s guidance, students formed small groups to explore. The first group focused on “health monitoring,” using smart wristbands to collect heart rate and blood pressure data from the elderly, employing machine learning models to identify abnormal fluctuations (such as sudden increases in heart rate), and designing a voice reminder function; the second group concentrated on “daily companionship,” training a voice recognition model so that the AI assistant could understand the elderly’s dialect commands (such as “Help me call my son” or “What’s the weather today”), and could generate a “daily mood report” based on the elderly’s chat records; the third group considered “safety protection,” combining cameras and image recognition technology to design a “fall detection” function—when an elderly person falls, the system automatically sends an alert to family members and dials the community service number.
During the project, students faced numerous challenges. For instance, the accuracy of dialect recognition was low, so they researched and learned about “speech feature extraction”; regarding the privacy protection of health data, they decided to use local storage + encrypted transmission; for the false alarm rate of fall detection (such as when an elderly person squats to pick something up being misjudged as falling), they optimized the model by adding a judgment condition of “action duration.” Ultimately, students not only completed the design of three functional modules but also conducted field tests in the community, collecting feedback from the elderly and iterating improvements. One student involved in the project summarized, “I used to think AI was far from life, but now I realize that as long as we observe needs with care, we can also use technology to solve real problems.”
The core of this stage is “openness” and “critique.” Teachers need to reduce the limitations of “standard answers” and encourage students to pose questions from different perspectives; at the same time, they should guide them to consider the “boundaries” of technology—such as whether the AI assistant system might lead to dependency among the elderly, or whether data collection might infringe on privacy. Through such inquiries, students can not only master AI technology but also learn to view technology from a rational perspective, growing into “empathetic technology users.”
4. Conclusion: Scientifically Advancing AI Education is Essentially “Student-Centered” Education
From the “perceptual interest” of lower primary grades to the “understanding principles” of upper grades, and then to the “application and innovation” of middle school, the curriculum design of AI education always revolves around one core—students’ cognitive development principles. While students at different grade levels may have varying cognitive development levels, this is not a simple “age segmentation” but a deep adaptation to students’ thinking characteristics and learning needs.
For teachers, scientifically advancing AI education requires letting go of the anxiety of “technological worship” and returning to the essence of education. Knowledge should be conveyed in ways that students can understand, motivation should be sparked by questions that interest them, and abilities should be cultivated through real challenges. When curriculum design resonates with cognitive development, AI education will no longer be an “extra burden” but will become the spark that ignites students’ innovative thinking, helping them better adapt to the future.
Author: Guo Weitong, Associate Professor, College of Educational Technology, Northwest Normal University
