82% of rural Chinese children between ages 6 and 36 months show developmental delays. Only 5% of elite university students come from rural backgrounds. These are the numbers China's Ministry of Education cited when it declared 2025 the start of "intelligentizing" education nationwide.
The plan sounds reasonable on paper: use AI to close a massive education gap. In practice, what's rolling out across Chinese schools looks less like an education initiative and more like a surveillance deployment with a textbook justification.
Smart Cameras in Every Classroom
The program spans four areas: teacher assistance with grading and lesson planning, rural school improvement, disability accommodations, and the one that should raise eyebrows - student biometric monitoring.
At Zhongguancun No. 3 Primary School in Beijing, AI systems analyze student classroom performance, homework completion, and emotional states. In Chengdu, Shuanglin Primary School has installed smart cameras that score student attention and participation in real time. Shenzhen's Bao'an Foreign Languages School monitors facial expressions during lectures.
The most striking example comes from Wuhan. Guanggu No. 9 Primary School ran AI psychological assessments on over 800 students, flagging 15 with "emotional issues" and 3 with "serious problems." There's no public information about what happens to students after they're flagged, or what safeguards exist against false positives.
The iFlytek Problem
Two companies dominate the rollout. iFlytek charged a single school roughly $254,000 for its system and landed a contract in Bengbu city worth approximately $220 million, covering 875 schools and 400,000 students. That works out to about $117 per student annually. Squirrel AI takes a different approach, offering 1.5 to 2 years of free access before subscription fees kick in.
Both models have problems. Previous technology rollouts in Chinese schools were plagued by predatory device-purchase schemes that pushed costs onto families. And the data privacy track record is poor: at least 63% of primary and secondary students have already experienced spam calls from data leaks tied to school systems.
Adding AI-powered biometric data to that leaky infrastructure should concern anyone paying attention.
Structural Problems Don't Have Software Fixes
The deeper issue is that China's education gap is structural, not technological. The hukou system (internal residency permits) restricts rural families from accessing urban schools. Only 6.8% of Chinese teachers held senior status in 2023, compared to 26% in U.S. schools. About 85% of education spending falls on local governments, and the Ministry's white paper sets priorities without providing funding.
Rural areas face a shortfall of 8.5 million computers against baseline requirements. You can't run AI classroom monitoring without computers in the first place.
Meanwhile, boarding school students - the very population this initiative targets - already show depression rates of 17.6%, with 8.4% exhibiting suicidal tendencies. Putting cameras on depressed teenagers and feeding their emotional data into AI systems is a choice that deserves more scrutiny than it's getting.
The technology exists to help teachers grade papers faster and generate lesson plans. Those are genuinely useful applications. But wrapping student surveillance in the language of educational equity makes the whole program harder to evaluate honestly, and that seems to be the point.