China builds AI vertical in education: from kindergarten to university (AI column)
Forbes. (12.01.2026) did analysisHow China makes AI a mass educational tool According to the engineer / CTO: this is a platform level program with regulation and KPI. Actually, the theses of the article are about such
What measures are being taken 1. Strategy and road map
- "Development plan of AI of the new generation" (2017) landmarked 2020 → 2025 → 2030. Education is a key training ground: after Education Informatization 2.0» (2019) Dozens of documents on AI-in-education; the goal is to 2030 AI is an everyday tool in all schools. 2. Scaling through pilots
- In February. 2024 PRC Minobra chose 184 Pilot schools/platforms for AI education (reference practices for replication). 3. AI from the diaper: content + tools Kindergarten: one-time classes / games with robot assistant teachers. Elementary school: textbooks where AI is a “wonderful friend”, then simple experiments / cases + ethics. Middle / Senior: Virtual Interviewers (Example: "Anna" for English)personalized DZ, autocheck part of the work. Separately: AI support for learning for children with ASD (Adaptation of tempo by emotion/state). 4. Graduate School: Staff Expansion + BigTech Participation
- In the list of bachelor's directions (04.2025) - 845 AI-related tracks (29 new)near 620 Universities offer a bachelor degree in Artificial Intelligence. BigTech opens laboratories/institutions and academies (Alibaba, ByteDance, Tencent, Huawei, etc.). 5. Regulation and Academic Integrity
- Chinese Academy of Sciences (09.2024): where allow for AI (search/structuring, editing, translation, graphs) - and what is prohibited; labeling/declaration and selection of "registered" services are required. Universities introduce local rules (Example: Fudan - only "non-creative" tasks, with the approval of the head).
What you see in the result n (figure)
- 1half 2025: 515 Millions of Gene AI Users (36,5percent).
- 83% in China believe that AI products/services have more advantages than disadvantages. Registered AI development: 117 (04.2024) → 801 (08.2025), +585%. Implementation is uneven across regions: “just digitalization” remains in places.
- Auto check is not perfect: in 2024 Zuoyebang on the tasks of the gaokao level dialed4/5.
The reverse side AI plagiarism and the breakup of “home vs. knowledge” (High scores for written, weak oral). The risk of unverified/generated information in science and the decline of critical thinking – hence the rise of rules and controls.
How to Repeat Outside of China Repeatably Pilots + measurable effect metrics (academic performance, teacher load, involvement). Age-related AI-curriculum + ethics/verification. AI is routine, but human-in-the-loop is at critical points. Policy “what can / can not” + change the format of the assessment (oral/practice/projects).
Harder. Super-scale and centralized model speed + unified infrastructure/regulatory.
- "Imposition of the standard" (For example, mandatory local/registered services).
#AI #ML #Education #Kids #Software #ML #Economics