China increasingly treats artificial intelligence not merely as a technology sector, but as part of its national education infrastructure. By 2030, the country aims to build a comprehensive AI education system spanning primary schools, secondary schools, universities, vocational training and lifelong learning. Almost 200 million pupils in primary and secondary education could ultimately be affected. Reuters describes the ambition as creating an “AI-native generation” — a generation for whom working with artificial intelligence becomes as natural as using search engines, smartphones or digital learning platforms is today.
The phrase “AI-native generation” is more of a journalistic description than the title of an official Chinese government programme. The policy direction behind it, however, is clear. In April 2026, five major Chinese authorities, including the Ministry of Education, the National Development and Reform Commission and the Ministry of Industry and Information Technology, jointly published the “AI + Education” action plan. By 2030, AI is expected to be deeply embedded throughout the education system, with a long-term structure for building AI literacy across society.
China is therefore taking a much broader approach than many companies that are still debating whether employees should be allowed to use ChatGPT, Copilot or Claude.
China is trying to turn AI literacy into a basic skill from school onwards.
AI Education Is Intended to Start in Primary School
The development did not begin in 2026.
As early as the end of 2024, China’s Ministry of Education published guidelines calling for artificial intelligence to be taught systematically in primary and secondary schools. The content differs according to age group.
In the first years of primary school, children are expected to encounter AI in a playful and accessible way.
Later, they should learn how AI systems work and how to use them.
At secondary-school level, pupils are increasingly expected to develop their own projects and engage with more advanced applications.
China is therefore deliberately avoiding an approach in which children merely learn how to operate a chatbot.
The goal is a staged model of AI literacy.
Younger children should understand that intelligent systems exist and what they can do. Older pupils should learn to use those systems critically. At higher levels, the focus shifts increasingly towards development, problem-solving and practical projects.
By 2030, AI Is Supposed to Run Through Every Level of Education
The 2026 action plan goes considerably further.
AI is to be fully integrated into local curricula. Schools are expected to offer dedicated AI courses while also using the technology across other subjects.
At universities, artificial intelligence is intended to become a general foundation subject.
That means AI is not supposed to be understood only by computer scientists.
Students in business, medicine, engineering, journalism and other disciplines are also expected to acquire basic AI skills. Universities are being encouraged to develop subject-specific teaching materials and connect AI with traditional academic fields.
The strategy therefore resembles, in some respects, the introduction of basic computer literacy several decades ago.
AI is not being treated as a specialist subject for a small group of experts.
It is being positioned as a general professional capability.
In Beijing, Every Pupil Receives AI Lessons
Beijing provides one example of what this strategy can look like in practice.
Primary and secondary school pupils there receive at least eight hours of artificial intelligence education per academic year.
For younger age groups, the emphasis is on experiencing and understanding the technology. Middle-school pupils focus more on practical applications. In upper secondary education, the emphasis increasingly shifts towards projects and innovation.
According to Beijing’s education authorities, AI was already being used in 87.7 per cent of schools by the end of 2025.
The city has also created teams of experts from universities, research institutes and technology companies to support schools with AI education.
China is therefore not only changing curricula.
It is also building an ecosystem connecting schools, universities, research and industry.
Shanghai Goes Even Further
Shanghai has developed an even more detailed curriculum.
In Years 4 and 7, pupils receive a dedicated subject on the fundamentals of artificial intelligence, with at least 30 teaching hours per year.
In later years, AI is intended to be integrated more deeply into computing, technology and other subjects.
At the same time, Shanghai is developing digital learning platforms, AI assistants for teachers and so-called dual-teacher models, in which teachers and AI systems jointly support learning.
More than 6,000 teachers in Shanghai were already using AI for lesson preparation, while more than 15,000 were using AI tools in teaching.
This makes it clear that the strategy is not only about teaching pupils about AI.
It is also about teaching with AI.
Pupils Are Building Apps and Robots
Reuters describes several examples of how far this development has already gone.
At the New Talent Academy in Beijing, pupils develop applications and build robots.
Other schools use AI systems to analyse lessons or generate teaching materials.
In rural regions such as Anhui, AI-generated educational content is also being used to help compensate for shortages of teachers and limited educational resources.
China is therefore pursuing two objectives at once.
In major cities, AI education is intended to develop talent for future technology industries.
In less affluent regions, the same technology is also being used to scale access to education.
AI Is Intended to Change the Role of Teachers Too
The strategy is not aimed only at pupils.
Teachers are also expected to build systematic AI skills.
Artificial intelligence is even expected to become part of teacher qualifications and certification. At the same time, teachers are being encouraged to use AI for lesson preparation, marking, personalised support and analysis of learning progress.
The Chinese action plan lists a broad range of possible applications.
AI can generate teaching materials.
It can manage homework.
It can answer questions.
It can mark assignments.
It can analyse classroom interaction.
And it can offer teachers suggestions for improving their teaching methods.
AI is therefore not being treated as an additional software tool.
It is intended to become part of the educational infrastructure.
Even Assessment Could Become More Automated
One of the most far-reaching areas is assessment.
China’s Ministry of Education describes scenarios in which artificial intelligence generates tasks, assembles exams, supports examination processes and automates parts of marking.
This could produce a much more detailed picture of how individual pupils are progressing.
At the same time, it raises significant questions about privacy, surveillance and algorithmic decision-making.
Once AI moves beyond supporting lessons and starts analysing performance or influencing educational decisions, the quality and transparency of these systems become critical.
AI Is Even Being Used in Physical Education
Examples from Chinese schools show how broadly the concept of “AI + Education” is being interpreted.
In Shenyang, computer-vision systems record sprint times and analyse movement patterns. Heart-rate sensors monitor physical effort during PE lessons. In calligraphy exercises, pupils can upload their work and receive automated feedback.
The boundary between educational software and continuous data collection is therefore becoming increasingly blurred.
For China, this broad digitalisation is part of a wider concept of so-called smart campuses.
At the 2026 World Digital Education Conference, corresponding standards were presented that included AI agents, multimodal systems and specialised education models as elements of future digital schools.
China Is Building Its Own Teaching Materials at the Same Time
To prevent AI education from depending entirely on individual schools, China is also developing centralised resources.
In May 2026, the Ministry of Education published a multi-part AI Literacy Education Handbook.
There are different versions for different age groups.
Younger children learn through stories.
Older primary pupils work on simple problem-solving tasks.
At secondary-school level, technical concepts are explained in increasing depth.
For older students, the focus shifts towards interdisciplinary applications and current developments.
There is also a course series called “Ten Lectures and 100 Questions on AI Literacy”.
The curriculum is not limited to technology.
Ethics, safety and responsible use are included as well.
The Key Term Is AI Literacy
That is where the broader strategic difference becomes clear.
China is not simply discussing how many people use a particular chatbot.
The goal is AI literacy.
That includes being able to use AI.
But it is equally important to understand when AI may be wrong.
One primary school in Shenzhen, for example, explicitly teaches children to question the answers of large language models and to verify information using multiple sources.
That is likely to become one of the most important skills of all.
Anyone who merely learns how to make AI produce results as quickly as possible is unlikely to retain a lasting advantage.
Someone who understands how to structure tasks, verify answers, choose suitable models and combine human judgement with machine assistance develops a far more durable capability.
That Is Different from Prompt Engineering
In many companies, early AI training was often treated as synonymous with prompt engineering.
Employees learned how to formulate better instructions for ChatGPT or Copilot.
That was a sensible starting point.
In the long term, however, it is too narrow.
Models are becoming better at interpreting imperfect instructions on their own. Natural-language interaction is becoming easier. Agents are taking over more intermediate planning steps autonomously.
As a result, the precise wording of a prompt becomes relatively less important.
Skills such as problem identification, task decomposition, source evaluation, data literacy, critical thinking and the supervision of automated processes become more valuable.
That is why the Chinese approach is strategically interesting.
The technology may change.
The ability to work effectively with intelligent systems remains.
China Is Trying to Solve Tomorrow’s Workforce Problem Today
There is clearly an industrial-policy dimension behind the education strategy.
China wants to become a global leader in areas such as artificial intelligence, robotics, semiconductors, intelligent manufacturing and autonomous systems.
Data centres and large language models alone are not enough to achieve that.
The country also needs millions of people who can develop these systems, apply them and integrate them into existing industries.
The education plan therefore explicitly connects schools, universities, vocational training and lifelong learning with the country’s broader economic transformation.
AI education becomes part of industrial policy.
The Scale Is Enormous
China has close to 200 million pupils in primary and secondary education.
If a country of that size systematically integrates AI education, it creates a potentially enormous future user base.
Reuters identified at least 31 million US dollars in publicly traceable government spending on AI education in recent procurement projects alone. The real figure is likely to be considerably higher because many local projects and investments were not fully captured in that analysis.
The economic effects extend beyond schools.
Providers of learning platforms, chips, robotics, educational software and AI models gain access to a vast domestic market.
Education policy therefore becomes technology policy as well.
China Still Does Not Have a Single Unified Standard
The strategy appears highly centralised.
Implementation is much less uniform.
Reuters reports significant regional differences.
Some wealthy cities have modern laboratories, robots and specialist teachers.
Other schools have far fewer resources.
There is also no single national definition of what an “AI-literate” pupil should be able to do.
Nor is there yet a nationwide standardised system for assessing AI literacy.
China therefore has a very clear direction.
But it does not yet have a fully standardised AI education system.
There Is Also a Fundamental Educational Problem
The more AI supports pupils, the harder one central question becomes:
When does AI support learning, and when does it replace learning?
A chatbot can explain a maths problem.
But it can also simply solve it.
It can help improve an essay.
But it can also write the whole essay.
Short-term performance may therefore improve without the underlying capability improving at all.
Research points to precisely this tension. AI support can improve homework performance in the short term. At the same time, there is evidence that intensive use under certain conditions may negatively affect exam performance, memory, independent thinking and analytical ability.
The existing research does not yet support simple causal conclusions.
Good AI Education Must Preserve Friction
Some Chinese teachers are already responding to this problem.
Pupils are expected to attempt tasks themselves first.
Only afterwards are they allowed to use AI.
In some cases, they must then explain what they have learned without AI assistance.
That approach may prove crucial in the long run.
An education system should not be optimised for pupils to produce correct answers with AI as quickly as possible.
It should be optimised for pupils to think better with AI.
That is a very different objective.
Europe Is Still Often Debating Restrictions
This is where the comparison with Europe becomes particularly interesting.
In many European schools, the debate continues to focus heavily on risk.
Should pupils be allowed to use ChatGPT for homework?
How can AI-generated text be detected?
How can cheating be prevented?
How should personal data be protected?
These are important questions.
China is asking an additional one:
What skills will pupils need if AI becomes a normal part of almost every profession within the next decade?
That perspective inevitably leads to a different strategy.
The Same Question Matters for Businesses
The real lesson therefore goes beyond schools.
Many companies currently follow a logic similar to that of education systems treating AI merely as a new tool.
They buy Copilot licences.
They offer a prompt-training session.
Then they hope for productivity gains.
That is unlikely to be enough in the long term.
Companies need an AI literacy strategy too.
Employees need to understand which tasks they can delegate.
They need to be able to verify results.
They need to know what data may be used.
They need to know how to work with agents.
And they need to recognise when an entire process should be redesigned rather than simply accelerating one existing step with AI.
A One-Off AI Training Session Is No Longer Enough
This is precisely why China’s education approach is relevant to businesses.
AI is not treated as something learned in a single training session.
It is taught over many years and at different levels of development.
The same principle can be transferred to organisations.
Beginners first need basic AI literacy.
More advanced users need role-specific applications.
Some employees will build agents and automations.
Executives, meanwhile, need skills in governance, organisational design and strategic evaluation.
A single “ChatGPT training course” therefore becomes a permanent learning model.
The AI Race May Be Decided Differently Than Expected
The global AI debate currently focuses heavily on models, chips and data centres.
Who has the most powerful GPUs?
Who trains the best model?
Who builds the largest data centres?
Those factors remain important.
But in the long term, another factor may prove just as decisive:
How many people can use AI productively?
A country can buy access to a frontier model.
A population with high levels of AI literacy cannot be created within a few months.
That is the long-term logic behind China’s strategy.
The Goal Is Not for Every Child to Become an AI Developer
It would be wrong, however, to interpret China’s plan as an attempt to turn every pupil into an AI engineer.
The objective is broader.
A relatively small group will develop models and systems.
A much larger group is expected to learn how to work with them.
The analogy is similar to computers.
Not everyone needs to design a processor or develop an operating system.
But almost every modern job assumes a basic ability to use digital tools.
China is betting that artificial intelligence will become the next such foundational technology.
The Biggest Advantage May Be Cultural
If children experience AI as a normal tool from primary school onwards, their attitude towards the technology is likely to change over time as well.
They will not have to encounter it for the first time in the workplace.
A 25-year-old entering the workforce in 2040 could already have more than a decade of experience working with AI systems.
That does not automatically mean they will make better decisions.
But the barrier to adoption disappears.
That is the core idea behind an AI-native generation.
Not people who blindly trust every AI system.
But people for whom working alongside intelligent systems is entirely normal.
China’s Strategy Should Not Be Romanticised
The speed and scale of implementation also come with serious risks.
AI-supported classroom analysis can quickly turn into continuous surveillance.
Health data, performance data and behavioural data about children are among the most sensitive categories of information.
If such data is collected centrally and assessed automatically, major questions arise around privacy, transparency and individual autonomy.
China’s strongly state-coordinated education system also means that technological objectives are closely tied to broader political education goals.
Speed of implementation should therefore not automatically be equated with quality.
China is effectively conducting a vast social experiment.
Its outcome remains uncertain.
Even So, China Is Asking Europe an Uncomfortable Question
The most important lesson is therefore not that Europe should copy China’s education system.
It should not.
The more important question is:
Does Europe have an equally long-term strategy for AI literacy at all?
Children starting school today will in many cases not enter the workforce fully until the late 2030s.
By then, AI agents, robotics and automated knowledge work are likely to be far more advanced than they are today.
An education system designed solely around the way people work in 2026 will therefore fail to prepare them adequately.
China Is Already Thinking in Generational Terms
That is what makes the Chinese strategy notable.
While companies debate which model to use and schools debate whether ChatGPT should be allowed for homework, China is already planning towards 2030 for an AI literacy infrastructure spanning the entire education system.
The technology will change several times before then.
Many of today’s models will probably no longer exist in their current form.
Many current tools will disappear.
But the fundamental ability to work with intelligent systems is likely to remain.
The global AI race may therefore ultimately be decided not only by which country develops the best models.
But by which country produces the largest number of people who know what to do with them.

