One Homeroom Teacher, the Rest AI: China Debates a Classroom Without Subject Teachers

Few jobs look as software-replaceable as lecturing, and this week a Chinese tech commentator said so in public. The blogger behind the handle "Third Uncle Who Drinks Black Tea" (喝红茶的三叔) posted that, having been forced to learn the basics of AI, he found the technology in his own word "terrifying": if development goes unchecked, most white-collar work will transform beyond recognition within a decade. His test case was the school. A class, he argued, needs one human — the homeroom teacher (班主任), the staff member who manages daily life, discipline and morale. Every subject teacher could be replaced by AI: delivering lectures, writing and grading tests, tutoring each student individually, "iterating to each student's level, without any strain." In a post liked more than 3,000 times, he added: "I genuinely don't know what society will look like ten or twenty years from now."

The rebuttal came fast, and it was not about capability. Another blogger, Song Xue, granted the premise entirely: yes, AI can teach the material. "But a child's emotions, character, and values — a machine cannot empathize," she wrote. "When children quarrel, when they fail and need comforting — that warmth between people is what AI cannot replace" (人与人之间的温度，AI替代不了). The division of labor she defended is distinctively Chinese: the homeroom teacher, or banzhuren, is the one adult assigned to a class's whole person — the one who calls home, notices the slump, handles the crisis. Both posters agree the machine can do the teaching. The argument is over what else a school has been quietly doing all along.

What lifts the exchange above punditry is the timing. On Zhihu, the question absorbing China's science-education world right now is how the 2026 National Physics Olympiad saw "large-scale AI-assisted answering" — not just in online practice but in the in-person semi-final and the offline training camps of the country's biggest prep brands, with one province's top known theory score now under scrutiny. Universities are tightening rules on generative AI in coursework in the same season. And 25 Fields Medalists' warning that AI is damaging how mathematics itself is learned gave Chinese teachers a version of the same worry at the highest level: the tool that can teach can also do the learning, and the learner is the one who loses.

There is a colder argument for the AI classroom that its critics concede: China's real shortage. Rural schools run on substitute teachers; a model that can hold a class of forty has equity appeal no reform has matched. But the same argument cuts at one of the country's last mass employers of graduates, in a year when youth unemployment anxiety remains raw. Policy has so far left the question open — no national ruling governs AI in the classroom — which is why one blogger's throwaway projection could tug at the profession's future. The debate this week ended where it began: the machine is ready to teach, and the country is arguing about the one human it wants to keep in the room.