AI News

The ragtag robot troupes flooding into primary and secondary schools

36Kr · AIUpdated Nov 16 18:42

AI summary

During the spring 2026 school-opening season, humanoid robots and robot dogs descended in droves on primary and secondary school campuses across the country, but most were outside hired performers costing several thousand yuan per day for two-day stints, not school assets. These robots, still in their "toddler stage," moved slowly, avoided obstacles sluggishly, and even malfunctioned by kicking out of control or clutching students, ultimately ending up as decorations in the corridors. In May 2026, the Beijing Economic-Technological Development Area delivered 20 "retired" humanoid robots to five secondary schools for secondary development, seen as a reference model for bringing technology into the curriculum structure.

Body · AI translation

The high school my cousin attends introduced two humanoid robots last fall—one stationed in the teaching building lobby as a greeter, the other performing a dance at the science festival. The school's official WeChat account posted an article with a title boasting "tech vibes maxed out," showing photos of the robots draped in red sashes, surrounded by a ring of curious faces. Three months later, my cousin's assessment of these two metal contraptions had shrunk to a single sentence: "Dumb and clumsy—slow-moving and always in the way."

The hallway isn't that wide, and during breaks hundreds of students pour out. The robot stands rooted at the corner, moving sluggishly and reacting slowly to obstacles. At first the students gathered around to watch, but later they switched to walking around it. "During peak hours, we're the ones dodging it—it really eats into our time racing for food," my cousin said. "Once a boy was running so fast he almost crashed into it, and the robot just stiffly shifted a little."

There's something absurd about this: a machine that was supposed to deliver "tech enlightenment" turns out to have its most reliable function in daily campus life be creating traffic obstacles. It didn't help students understand the technical logic of embodied intelligence, nor did it become a lab assistant in tech classes—instead it became a "troublemaker" that has to be carefully preserved.

If you pull your gaze back from this one hallway, you'll find similar stories playing out densely over the past year. Robots are flooding into primary and secondary school campuses in the roles of "new classmates," "greeters," and "guest performers."

The next school, the next ceremony, the same batch of machines, the same set of moves... So why exactly are robots crowding into campuses? Is bringing smart technology into schools actually a meaningful thing?

Rented Tech Vibes

In the spring 2026 school-opening season, humanoid robots and robot dogs became practically standard equipment for opening ceremonies at primary and secondary schools across the country.

Many schools brought in "the same models as seen on the Spring Festival Gala." On the playground of Lucheng Foreign Language Primary School, a humanoid robot served as a "pace-setter" running with students; at the gate of Songgang No. 2 Primary School, a robot dog bowed and shook hands, surrounded by children posing for photos; one school in Longgang District brought in seven "robot guests" at once—one writing auspicious characters, one doing kung fu, one breakdancing, each with its own job. At the opening ceremony of a rural primary school in Mianyang, a robot draped in a red sash greeted people at the school gate, then sang "Running" in chorus with a teacher as children clustered around it walking into the classroom, "observing curiously and communicating excitedly." Shanghai, Foshan, Jinan, Nanjing—similar scenes repeated themselves over and over within the same week.

The spectacle was lively, but the sourcing was mostly slapdash. Most of these robots were rented reinforcements, not school assets. A humanoid robot capable of dancing and interactive movements rents for several thousand yuan per day; schools usually need to rent two or more, for a two-day period—one day of rehearsal, one day of the actual performance—to complete a set of waving, bowing, and dancing moves at the opening ceremony before being loaded onto a truck and hauled away. The cost of two machines for two days is enough to cover a semester's club activity budget for an ordinary primary school. For schools in remote areas, that sum weighs even heavier.

Still, students standing in the audience watching a robot dance and watching a robot dance at a science museum aren't fundamentally different experiences. The science museum even offers richer exhibits, with explanations, interaction, and no need to bear the risks of equipment transport and on-site debugging. Renting robots to perform at school is, in essence, not much different in educational logic from hiring a performance troupe to liven things up.

Some principals will defend it: "The intellectual awakening brought by immersive experience can't be measured in money. A rural child who has never seen a robot before, reaching out to touch a robotic arm for the first time—that tactile sensation might plant a seed." That makes sense, but for a seed to sprout, it needs soil afterward.

If, once the ceremony ends, the robot is hauled away and everything in the classroom goes on as before, that seed most likely stops at the level of having been touched. The two robots allocated to my cousin's school by the district—which, lacking supporting curricula and teachers, ultimately degenerated into hallway decorations—are a ready-made example. If even allocated ones end up like this, rented ones need hardly be mentioned.

This is a structural haste born of chasing the trend: the technology isn't mature yet, the curricula aren't in place yet, the teachers aren't prepared yet—but the ceremony can't wait.

A New Classmate That Hasn't Learned to Walk Yet

My cousin thinks the robots are "dumb and clumsy," a view highly consistent with the state of the industry.

In April 2026, at the opening ceremony of a university sports meet in Xinjiang, a humanoid robot participating in the performance suddenly lost control, and after falling to the ground began kicking and flailing wildly at the students around it, before finally being carried off the scene by staff. The same month, at a campus event in Xi'an, a performance robot malfunctioned and "hugged" the girl dancing alongside it.

When these videos circulated on social platforms, the most common comment was: "Thank goodness nothing happened."

These incidents are no accident. The technical stage humanoid robots are currently at is accurately described as "the toddler phase." At a training ground in Beijing's Shougang Park that the media calls a "robot kindergarten," fifteen robots of different forms are divided into classes by ability, trained level by level from hardware-software adaptation to body perception to social collaboration. Specifically, the robots stumble when they walk, pause, get stuck, occasionally fall down and collide with each other, and staff have to "carry" them out of the area from time to time. One of the admission criteria is being tough and durable, plus fitting a soft outer shell "to reduce injury from falls or human-robot interaction."

In other words, at the industry's most cutting-edge training ground, humanoid robots are still iterating on something as basic as walking steadily through repeated falls. Putting these machines—still at the "kindergarten" stage—among thousands of primary and secondary school students and expecting them to handle greeting, guiding, and performing is itself a mismatch.

The deeper problem is that this mismatch is concealed by the rhetoric of "tech feel." A robot that dances flawlessly at an opening ceremony is displaying choreographed, repeatedly debugged successful motions. It does not expose failures, does not show the debugging process, and does not let students see "why machines are dumb." What students receive is a finished product, not a technological object that can be understood, questioned, and modified.

And this slapdash performance approach is exactly the opposite of the goals of science and technology education.

What science and technology education wants students to gain, in the end, is an understanding of technology—knowing how a machine moves, why it fails, and where it can be changed. But the logic of performance is that no mishaps are allowed; if something goes wrong, the robot is quickly carried away, and the anomaly becomes an accident to be covered up. Students are also positioned as spectators standing at a safe distance, clapping or being startled, rather than getting hands-on to take things apart, ask questions, and modify them.

And this, in turn, pushes students further and further away from technology.

The divide over bringing intelligence into schools

If "whether intelligence is worth introducing into schools" is to be treated as a question that can be judged, the criterion is not whether it is novel enough or photogenic enough, but whether it has entered the curriculum structure and whether it gives students room to "tinker" with it. A robot that students can write code for, adjust parameters on, and repeatedly trial and error with—even if it walks unsteadily—its clumsiness itself is teaching material. A robot that can only dance flawlessly on stage and then be packed into a box and hauled away, no matter how smooth it is, is merely a consumer product.

As for embodied intelligence, Beijing E-Town provided a sample worth referencing in May 2026. They delivered twenty humanoid robots "retired" from robot soccer matches to five middle schools in the district, hoping to integrate them into school-based curricula and club activities in artificial intelligence and robotics, develop dedicated teaching modules, and encourage students to carry out project-based learning and research around them. A vice principal of one of the receiving schools once said: "We look forward to these humanoid robots, under the guidance of frontline teachers, helping students with secondary development."

There is a clear dividing line between secondary development and watching a performance. The former means students have to write code, adjust parameters, face failure, and repeatedly trial and error. A robot that cannot walk steadily is a defect in a performance scenario, but it is material in a development scenario. The clumsier it is, the more teaching value there is in the students' debugging process. The latter requires only that students sit in the audience and clap at the appropriate moments.

This logic applies equally to intelligent technologies beyond robots. Large models are entering campuses faster than humanoid robots, and the risks are more hidden. Intelligent grading systems can free teachers from repetitive labor; intelligent Q&A can let students learn subject knowledge in a fun, one-on-one way; and fun intelligent games can also spark students' interest in learning and make learning enjoyable.

It is worth noting that in these cases the role AI plays is not to complete tasks for students, but to free students from inefficient steps so they can do things more worth doing. Mechanical steps such as grading, data collection, and preliminary feedback are taken over by machines, and the attention saved by students and teachers is directed toward design, questioning, and iteration.

However, most of the robots currently flooding into primary and secondary schools are demonstrative technologies—lively as they are, they have not truly landed on educational value. What schools should really introduce is technology that can enter the curriculum structure and allow failure and retries.

Twenty thousand yuan for a performance of tech feel buys a tweet and a few photos; twenty thousand yuan for a robot kit buys dozens of failures and dozens of retries over a semester. Once the former sum is spent, it is over; once the latter sum is spent, learning has only just begun.

This article is from the WeChat public account "Naojiti" (ID: unity007), author: Shanhu, published by 36Kr with authorization.

Source

36Kr · AI