Welcome to the special edition of China Scholar Insights!
When people discuss the global AI race, the conversation usually revolves around chips, computing power, frontier models, and investment. These are all important indicators of technological strength. But there is another question that receives much less attention: What happens after powerful AI systems are built?
A country may develop advanced models, but their broader economic and social impact still depends on whether businesses, workers, teachers, and ordinary users can actually understand and use them effectively. This raises an interesting question: Could AI literacy become another important dimension of national AI competitiveness?
I found this question worth thinking about. Is AI competition ultimately determined by who builds the most powerful systems, or will it also depend on which societies are better able to spread AI capabilities across the economy and everyday life? China’s recent push to expand AI education and public AI literacy offers an interesting case through which to examine this question. These developments prompted Mingyin Xie and me to take a closer look at China’s emerging effort to build AI literacy at scale and write this article.
Questions, comments, or criticisms are welcome at sch0625@gmail.com
When observers discuss the global race in artificial intelligence, the same indicators tend to dominate the conversation: access to advanced chips, computing power, the performance of frontier models, investment, and the size of a country’s pool of AI researchers.
These measures matter. They tell us a great deal about which countries are capable of pushing the technological frontier forward. But they tell us much less about what happens after an advanced model is built.
Can teachers use AI effectively in classrooms? Can workers adapt as AI reshapes their jobs? Can small businesses and public institutions make productive use of the technology without dedicated technical teams? And can ordinary users recognize its limitations rather than simply trust whatever an AI system produces?
China has long included public AI education in its national strategy. Recent policies are making that ambition more concrete by linking education and training more closely to the widespread use of AI.
Why AI Literacy Matters
The economic impact of AI ultimately depends not only on innovation, but also on adoption. A highly capable model has limited wider value if its use remains concentrated among technology companies and specialists. Its effects become much larger when teachers, engineers, researchers, manufacturers, civil servants and small businesses can incorporate it into their work.
This points to two dimensions of national AI capacity. The first is frontier capability: the ability to develop increasingly powerful AI systems. The second is diffusion capacity: the ability to spread those systems through the economy and society and enable people to use them effectively.
Frontier capability can be tracked through chip access, computing resources and model benchmarks. Diffusion requires a different set of questions: who can access these systems, which tasks they can use them for, and whether their use improves outcomes. Education and workforce skills help connect technical capability to practical value.
For China, diffusion has become an explicit policy priority. The State Council’s 2025 “AI+” initiative sets out measures to expand AI use in science, industry, consumption, public services and governance. It also calls for improving AI literacy and skills among workers, expanding training and assessing employment risks. Education and skills are therefore part of the policy’s approach to adoption.
AI literacy combines a basic understanding of how AI works with the ability to use it, evaluate its outputs and make responsible decisions about its use. For an ordinary user, that might mean checking a generated claim, protecting sensitive information or recognizing when a task requires human expertise. These capabilities form part of the human infrastructure needed for widespread AI adoption.
How China Is Building AI Literacy at Scale
China’s approach has become much more systematic over the past year.
In April 2026, the Ministry of Education and four other government departments issued an “AI + Education” Action Plan. It sets a target for 2030: an AI education system spanning all stages of formal education and broader public learning, supported by a sustained mechanism for developing AI literacy across society.
The intended audience extends well beyond future AI scientists and software engineers. At the school level, the plan calls for AI education to be integrated more systematically into curricula. At universities, the plan calls for AI to become part of the foundational curriculum for students across disciplines, with teaching materials adapted to different fields of study. Vocational education is expected to connect AI training more closely with changing workplace needs, while universities and open-learning institutions are encouraged to provide AI literacy and skills courses for wider groups of learners. It also calls for AI knowledge to be incorporated into teacher qualification examinations and certification.
The effort already has substantial reach. By August 2026, China’s national smart-education platform had gathered more than 1,000 AI courses spanning basic, vocational and higher education as well as lifelong learning. Ministry-organized training had reached 250,000 school leaders and local education bureau heads, while AI-literacy training had covered more than 310,000 primary and secondary school teachers in central and western China.
The agenda extends beyond formal education. In May 2026, In May 2026, four government bodies made public AI literacy one of six priorities in their annual digital-literacy work programme. The same month, an “AI Literacy Reference Framework” was released at the China Internet Civilization Conference, signaling an effort to think about AI competence at the level of society rather than schools alone.
China is part of a wider international shift. In April 2025, the United States issued an executive order promoting AI education, teacher training and apprenticeships, linking them to an “AI-ready workforce.” The July 2025 U.S. AI Action Plan also called for AI skills to be integrated into career and technical education, workforce training and other federally supported programmes.
In June 2026, the OECD and European Commission published an AI-literacy framework for primary and secondary education. It emphasizes understanding AI systems, evaluating their outputs and using them ethically and creatively. Across these initiatives, AI literacy is increasingly treated as a capability needed beyond the technology sector.
What deserves attention in China’s case is the attempt to connect a nationwide education agenda with industrial adoption, adult training and public digital literacy. AI literacy is being placed within a broader effort to expand AI use across the economy. Whether these policy connections translate into better learning and more effective workplace use remains an open question.
The push also reaches people already in the workforce. In August 2026, Guangdong’s human resources and social security department launched a province-wide AI training initiative. Running through January 2027, it targets at least 1,000 AI courses, 1,000 training activities and one million training participations. It offers a practical example of how the literacy agenda is extending beyond schools and universities.
The Promise and the Limits
Access to capable models does not automatically translate into effective use. Workers still need to understand where AI can improve existing tasks, how to evaluate its outputs and when human judgment remains necessary. The same applies to teachers, public institutions and smaller businesses that may lack dedicated technical teams. Broader AI literacy could therefore make it easier for AI to move beyond technology companies and specialists into everyday economic and social activity.
China’s national platforms, universities and vocational schools provide channels for delivering AI education at scale. Their practical value will depend on how well common resources are adapted to different learners and occupations. A factory technician, a teacher and a small-business owner need different forms of training. Connecting these needs to the wider “AI+” agenda will require continued cooperation between education providers and employers.
Large-scale adoption is only a starting point. Knowing how to write a prompt does not establish whether a user can evaluate the answer. The OECD’s Digital Education Outlook 2026 draws a related distinction: generative AI can improve students’ performance on a task without producing lasting learning gains. Educational benefits depend on how the technology is designed and used, including whether it supports students’ own reasoning.
As AI becomes easier to use, some of the most valuable skills may be precisely those that cannot simply be delegated to it: critical thinking, independent judgment, the ability to identify unreliable information and knowing when an automated system should not be trusted.
Recent Chinese guidance also addresses these concerns. An AI education ethics framework, released at the World Digital Education Conference in May 2026, emphasizes human responsibility and distinguishes between prohibited, restricted and encouraged uses. It provides guidance for educational practice; its significance will depend on how institutions translate those principles into everyday decisions.
Implementation may pose another challenge. Schools and communities do not have equal access to teachers, infrastructure and digital resources. Universities and large organizations may adapt faster than smaller institutions, rural schools or older workers. A strategy intended to narrow the AI skills gap could create new inequalities if the quality of access and training varies widely.
There is also a deeper question about what an “AI-literate” person should actually be able to do. If AI can write, translate, summarize, calculate and increasingly perform more complex cognitive tasks, education systems will have to decide which capabilities people should continue to develop independently. The goal cannot simply be to produce a society in which everyone uses AI.
China’s bet will be tested by what people can do after the training: whether teachers can improve learning, workers can adapt to changing tasks, and users can identify errors and make informed decisions. Course numbers and participation figures show reach. The stronger evidence will be improvements in performance, adaptability and independent judgment that endure as AI tools change.
Mingyin Xie: Research Associate at ChinAffairsplus.
Chenghao Sun: Founder and Editor-in-Chief at ChinAffairsplus; Associate Professor at Tsinghua University.
