Jiangning School explores "Suitable Education + AI" classroom integration: let technology step back, let students step forward

Enter a few prompts, and AI can automatically generate a piece of music or write out a set of experimental plans. As intelligent technology accelerates its entry into schools, will students be led by algorithms, or will they learn to explore more independently with the help of technology? At a teaching demonstration event recently held at Shanghai Jiangning School, several ordinary subject classes gave the answer: AI should not be a flashy demonstration used to "put on a show," but should become a lever that drives real thinking. It can help teachers diagnose learning conditions and simplify the complex, and it can also guide children to transform from passive "answer-givers" into active "truth-seekers." In the chemistry class, AI serves as an "assistant," and real problems become the fulcrum of the classroom. A seemingly simple interdisciplinary project of "making a small boat" has been followed up, practiced, and continuously optimized by the eighth-grade physics teaching and research group at Jiangning School for three consecutive years. "In the past, after students finished making a small boat and put it in the water, it would tilt as soon as an egg was placed on it, and they wanted to improve it but did not know where to start." Zhao Jiajun, a senior physics teacher, introduced that the teaching and research group collected students' after-class questionnaires and found that the problems everyone encountered were different: some boats swayed and needed added stability, some hoped to carry more heavy objects, and others wanted to solve the hidden risk of water entering through hull damage. However, "in the face of diverse improvement needs in the class, limited classroom time makes it difficult for teachers to tutor each student one by one." At this point, the teaching and research group did not directly stuff ready-made improvement plans to students, but instead developed an intelligent agent called "Boat Assistant." After obtaining materials from the intelligent agent, students need to extract technical cards themselves, compare shipbuilding techniques from different eras, and then optimize their own small boats by hand. To help everyone understand the physical principle of "watertight compartments" preventing sinking, the teaching and research group used AI to create a dynamic model of water entering the hull step by step, which corroborated the real demonstration experiment in class. Not only that, the teachers also set strict rules for this "AI assistant": sources must be strictly limited to professional books, official institutions, or museum websites; generated answers must cite sources; and if no basis can be found, it must directly say "I don't know"; the intelligent agent is only responsible for guiding students to think through questions and must never directly replace students in generating design plans. "AI helped us quickly build a prototype of a teaching tool, but scientific validity and applicability have always been checked by the teacher team," Zhao Jiajun said. The real fulcrum has always been in students' real learning problems. Human-computer interaction makes students' thinking more active. In the class "The Color-Changing Code of Sappanwood" taught by Pan Qinyue, a first-grade chemistry teacher, the starting point of inquiry came from a life situation: dyeing colored flags for the school sports meet. What color gradient would sappanwood extract undergo when encountering different pH levels? In class, students split into multiple groups to carry out experiments. The AI system quickly summarized each group's experimental conditions and test tube color photos in the background, making cross-group comparison clear at a glance. Relying on a virtual laboratory developed with AI, children can independently adjust solution concentration and acid-base conditions before actually doing hands-on work, intuitively simulate and correct errors on the screen, and then conduct practical trial dyeing. The whole class follows the main line of "raising questions—experimental inquiry—evidence analysis—transfer and application." The intervention of technology makes the relationship between abstract pH values and concentration clearly visible, and also allows the scientific thinking of controlling variables to take root naturally. In the music class, technology likewise stimulates students' artistic creativity. Under the guidance of music teacher Yu Dawei, students tried to collaboratively generate melodies on an AI platform, but people always firmly controlled the aesthetic initiative. Through comparative listening and analysis of rhythm and emotional expression, students continuously adjusted parameters and revised generated fragments, and finally sang their own campus melody in vocal chorus. Pay attention to "how to learn," and even more to "healthy growth." When technology enters daily life, how can children be prevented from becoming overly dependent on algorithms? "For education, the intelligent era is not only a question of technology application, but also a question of physical and mental health," admitted Qi Yue, the school's full-time psychology teacher. AI does have the functions of immediate positive feedback and highly realistic companionship, but while assisting learning, it may also bring hidden concerns such as mental inertia and withdrawal from real-life social interaction. For this reason, relying on a Shanghai educational science research project, the school adheres to the concept of "preventing disease before it occurs" and, starting from the four dimensions of algorithms, computing power, scenarios, and data, guides students to learn to identify information cocoons and protect real interpersonal communication through special courses, thematic group counseling, and the construction of psychological corners. "What we want to cultivate is masters of technology, so that children can maintain independent thinking and a sense of autonomy amid the digital wave." As a municipal-level AI experimental school, Wu Qinglin, principal of Jiangning School, said that the key bottleneck in bringing technology into the classroom is not hardware, but teachers. By building a four-stage progressive training system of "cognition popularization—tool practice—teaching integration—innovative design," the school has established a "seed teacher" mechanism to drive all teachers to grow from "users" of technology into "designers" of classroom learning. Wang Zhenfei, director of the Putuo District Education Bureau, said that at present, Putuo District is deeply cultivating the path of "suitable education + AI," and the participation rate of primary and secondary schools in the district in artificial intelligence education has exceeded 90%. "The yardstick for regional progress in digital transformation is definitely not simply how much equipment has been added to classrooms or how much software has been used, but whether technology can effectively help teachers assess learning conditions and optimize teaching, and whether it can stimulate students' inner drive to ask questions proactively and think deeply, truly transforming technological advantages into educational effectiveness."