China's $31 Million AI Classroom Push: Big Ambition, No Scoreboard
Reuters tallied $31 million in Chinese AI classroom tenders as provinces mandate AI hours and exams, yet no shared framework exists to measure whether students gain AI literacy.
By Amara Osei
4 min read

What's News
- Reuters documented at least $31 million in Chinese government AI classroom procurement over the past year from 100+ tender records.
- China targets universal AI literacy for nearly 200 million primary and secondary students by 2030 under an April 2026 action plan.
- A study of nearly 27,000 teenagers found AI use tied to an 18% rise in homework scores but a 20% drop in monthly exam scores.
- iResearch estimates 2025 education digitalisation spending at just under $83 billion, with generative AI at 46% of AI education procurement.
- Yunnan plans AI lesson analysis in 14,000 classrooms; Ningxia and Gansu will add AI literacy to entrance exams by 2027.
Chinese government bodies committed at least $31 million to AI classroom equipment over the past year, according to a Reuters review of more than 100 public tender records. The number is modest against China's education system, but it quantifies a national drive to make AI literacy universal among nearly 200 million primary and secondary students by 2030.
The procurement tally sits inside a weeks-long Reuters investigation combining school visits, interviews with more than a dozen educators, parents and experts, and analysis of over 100 academic papers and government documents. The picture that emerges: broad ambition, uneven execution, and no shared framework for measuring whether students actually become AI-literate.
How does Beijing deliver the goal?
Since 2024, the Ministry of Education has pushed schools to introduce AI courses, train teachers and launch pilot programmes. An April 2026 action plan from five national agencies formalised AI education at every stage by 2030, with security assessments and risk-response mechanisms for cheating and privacy leaks.
Nearly half of China's provinces and municipalities have issued binding implementation plans. The local mandates include:
- Shanghai: at least 30 hours of AI-skills teaching per grade each year
- Beijing: 100 AI pilot demonstration schools plus AI in student health monitoring
- Ningxia and Gansu: AI literacy in high-stakes entrance exams by 2027
- Guizhou: facial recognition and behavioural analysis in classrooms
- Yunnan: AI recording and analysing lessons in 14,000 classrooms
The scope runs well past coding lessons. Consultancy iResearch puts 2025 education digitalisation spending at just under $83 billion, with generative AI projects accounting for 46% of AI education procurement. The $31 million in documented tenders likely captures a fraction of real outlay, scattered across thousands of local budgets that nobody tallies centrally.
Two schools, same models
At the $39,000-a-year Beijing New Talent Academy, 13-year-old Peng Gaodi has built more than 200 apps using AI, including a homework-submission platform and a secondhand marketplace. The school runs its own digital innovation curriculum covering AI, robotics and programming, and has built a classroom AI model drawing on popular Chinese systems. Headmaster Liu Wei said only a small minority of students will become AI scientists; for the rest, the goal is making AI an everyday tool for learning and creating.
In rural Anhui, Funan County Qinqing High School uses the same underlying models — from Alibaba's Qwen, ByteDance and Z.AI — against a different problem: a shortage of qualified teachers. Its 4,000 pupils, mostly children of migrant workers, follow personalised lessons on tablets. AI agents pose questions rather than supply answers when students struggle. Headmaster Yang Tianliang credited the approach with lifting university admission rates from below 60% to 90%, figures Reuters could not independently verify. Yang estimated at least 200 other schools run similar systems.
Does the evidence support the rollout?
"Mature, widely adopted assessment rubrics for compulsory school-level AI competency are still being built," said Selena Guo of research firm China Policy. Without a standard, schools are experimenting without a shared scorecard.
The early research conflicts. A study of nearly 27,000 Chinese teenagers linked AI adoption to an 18% rise in homework scores but a 20% decline in monthly examination scores within six months, with the largest drops in social science subjects. A separate nationwide survey of more than 320,000 pupils found educational benefits alongside concerns about dependence and weaker independent thinking. Students using AI at least three times weekly scored lower on memory, comprehension, analytical and creative skills than less frequent users. Neither study establishes causation.
Ted Mo Chen, chief of staff at ed-tech startup ClassIn, told Reuters implementation is "all over the place" and that few schools he visited had integrated AI effectively into the curriculum. He warned children's thinking processes could be "completely bypassed" if schools allow reliance on AI for quick answers. Academic research reviewed by Reuters also suggested rural students may be particularly vulnerable to cognitive outsourcing and device addiction.
Who is building guardrails?
Educators and parents are improvising their own. At Beijing New Talent Academy, Peng said he tries to solve coding problems himself before asking AI for hints rather than complete answers. Beijing parent Rita Luan requires her 11-year-old daughter to spend 15 minutes on a hard maths problem before consulting ByteDance's Doubao chatbot — then solve the same problem days later without help. "The point is to make sure she actually digested the solution," Luan said.
The Ministry of Education did not respond to Reuters' request for comment. Until a national assessment framework exists, China's $31 million in documented procurement measures intent, not outcomes — a signal that the world's largest AI-in-education experiment is still flying without instruments.
Source: epyhia.com
More from Amara Osei
Show full bio
Senior reporter covering companies, deals and strategy in entrepreneurship, startups and venture funding at Business Bearings.
2,136 articles