Illustration of students working with a robotic arm in a classroom
World

CHINA WANTS 200 MILLION STUDENTS TO LEARN AI BY 2030

China aims to make AI literacy universal among schoolchildren by 2030. Here is what nearly 200 million students means for classrooms.

By PRESDA Editorial4 min readUpdated

Editorial illustration supplied for PRESDA. It depicts a conceptual classroom.

China wants AI literacy to reach its schoolchildren by 2030. The country counts nearly 200 million students in primary and secondary education, Reuters reported after visiting schools and reviewing provincial plans. The number gives the scale of the effort. It is an enrolment estimate, rather than a government tally of pupils who have completed a course. The April 2026 action plan seeks to integrate AI throughout education and establish a system for teaching it across age groups. Chinese Ministry of Education action plan Reuters classroom reporting

What the national plan asks schools to do

The plan covers primary and secondary schools, universities, vocational training and lifelong learning. The ministry describes a different priority at each level: basic literacy in schools, work across disciplines at universities, and practical skills in vocational education. This matters because recognising when a chatbot is unreliable is a different skill from training a model or using one to inspect manufactured goods. The policy also covers using AI in teaching and assessment, which is distinct from teaching students how AI works. Chinese Ministry of Education action plan Ministry of Education implementation briefing

National ambition and local examples
ScopeWhat it measures
Nearly 200 millionPrimary and secondary enrolment cited by Reuters
2030National integration target
30 hours a yearShanghai requirement per grade, as reported by Reuters

What AI literacy might look like in class

A lesson can start with a familiar question and two conflicting chatbot answers. Students can identify the claim each answer makes, look for corroborating evidence and explain which source they trust. Another class could test an image classifier with new examples and record which images it mislabels. Both exercises make students explain the result. Handing in a polished answer would reveal much less about whether they understood the system. This distinction matters when schools use the same tools to help teachers assess the work.

What Shanghai and Beijing show about implementation

Reuters found that nearly half of the provinces and municipalities it reviewed had rolled out binding plans. It reported that Shanghai requires at least 30 hours of AI instruction a year per grade. Beijing plans 100 pilot demonstration schools. One measure is a minimum amount of instructional time, while the other is a network of sites where approaches can be tried. Neither tells us what a student in a different locality receives, or whether an hour of instruction has the same content across schools. Reuters classroom reporting

Why teacher training matters as much as access

Buying a platform can be counted quickly. Preparing a teacher to choose a good exercise, spot errors and give meaningful feedback takes longer. Schools also need decisions about which tools pupils can use by age and how classroom work should be evaluated when AI assistance is allowed. Shared teaching material can make lessons easier to access. It cannot by itself provide reliable connectivity or local support to every school. The ministry’s plans emphasise institutional capacity and teacher development as well as digital services. Chinese Ministry of Education action plan Ministry of Education implementation briefing

What will tell us whether the plan is working?

The strongest evidence would show what students can do after the lessons: check an invented citation, explain why a prediction failed or complete a task without handing the judgment to a chatbot. Participation can be measured separately from proficiency. Reporting enrolment, course completion and demonstrated skills as distinct results would make comparisons clearer. Schools using AI to analyse student activity face another question about who may access the data and how long it is retained. The spread of tools such as Alibaba’s Qwen4-Max helps explain why verifying an AI answer is now a classroom skill. Chinese Ministry of Education action plan Reuters classroom reporting

How wide access could produce different outcomes

The same tool could support very different lessons. In one classroom, a teacher might ask pupils to explain why an answer is wrong after comparing it with their textbook. In another, pupils might use the tool mainly to draft a response. Both schools could report that AI was used, although only the first exercise directly tests verification. That is why an evaluation needs to distinguish exposure, frequency of use and the skill gained. It should also account for differences in equipment and teacher support, so changes in scores are not automatically credited to the software.

FAQ

Frequently Asked Questions

Does China have an official target of exactly 200 million students?

The number is Reuters’ approximate primary and secondary enrolment figure. The policy goal is broad AI literacy and integration by 2030.

Does each pupil take the same course?

Local plans vary. Shanghai’s reported hours should not be treated as a nationwide minimum.

#CHINA#World

PRESDA Dispatch

Stay Informed. Stay Aware.

A sharp briefing across AI, gaming, sport, business, world affairs, paparazzi, and lifestyle.