In China, instead of creating a new subject called “AI”, the country integrates AI content into existing subjects such as math, science, technology, and engineering. 

From primary school, students get familiarized with computational thinking. In secondary school, they are exposed to basic programming and data-driven problems. In high school, advanced content such as computer vision, chatbots, and machine learning models are piloted.

The government plays a central role in policymaking and coordinating resources nationwide. Tech firms provide software, materials, and educational technology support. All companies, from iFlytek to Baidu, have “AI for schools” programs. Meanwhile, top-tier universities are tasked with developing curricula, training teachers, and evaluating implementation quality.

The Chinese government has developed a national AI learning platform that allows students from all regions, including poor areas, to access the same content as students in Beijing or Shanghai. 

While China is following the top-down model, the US is carrying out the bottom-up reform, leveraging its decentralized education system for flexible experimentation. 

Together, with an open letter from over 250 CEOs to state governors, tech giants like Microsoft, Amazon, Meta, and NVIDIA have launched programs supporting public schools with free AI learning software, teacher training, equipment, and sample curricula.

Students interact with AI chatbots during math lessons, use computer vision to do biology experiments, and learn programming through AI-enabled games.

The federal government is stepping in with an “AI Education Task Force” to standardize curricula, connect initiatives, and ease regulatory barriers for private sector involvement. The Department of Education collaborates with states to develop open learning resource repositories, teacher training hubs, and pilot programs in underserved areas.

However, both China and the US face significant hurdles as AI permeates education, not just technical, but also social and ethical issues.

First, data privacy. When students use AI tutors, data about their learning behaviors, emotions, information processing speed, and even how they raise questions are collected. Without legal protection, companies can easily commercialize this data, or use it to tailor content to their own advantage.

Second, technological differentiation risk. In the US, the gap between rich school districts (usually urban areas) and poor school districts (rural, ethnic minorities) will widen without adequate federal investment. In China, the “AI tutor” model may be effective in areas with good infrastructure, but useless in areas without basic digitalization conditions.

Third, shaping thinking through algorithms. Students may unconsciously absorb the biases hidden in the algorithm. From there, education may lose its role in shaping independent thinking – the core of a democratic society.

What should Vietnam do?

Vietnam is at the starting point of AI education design. The question isn’t whether to emulate the US or China but to identify which approach suits Vietnam’s infrastructure, population, and teacher readiness.

Vietnam can learn from the Chinese model. Schools in Vietnam can integrate AI into existing teaching subjects instead of creating new subjects. 

The Ministry of Education and Training (MOET) should set a minimum competency framework for computational thinking and AI at each level of education. Building an open, shared national digital science repository will help reduce inequality between urban and rural areas, and lowland and mountainous areas.

The Ministry should define minimum computational thinking and AI competencies for each school level. A national open digital learning repository could reduce urban-rural and regional disparities.

The positive aspect from the US that Vietnam can consider is mobilizing the private sector to participate in the program on training teachers and providing education AI platforms.

Vietnam’s FPT, Viettel, VNPT, VNG, and CMC can play the role as the US’s Microsoft and NVIDIA do, not only in developing infrastructure, but also in developing open standard learning software. Mass teacher training via digital platforms, offering MOOC-style certifications, should be prioritized.

Vietnam should also consider establishing a national coordination center, which could be the “National AI Education Committee” to unify curricula, connect businesses, schools, and the state, and integrate national learning data. This body should avoid rigid bureaucracy, prioritizing transparency and adaptability.

Vietnam does not need to be a “copycat” of anyone. The most important thing is to start now: build an integrated AI program from primary school level, train teachers widely, make learning devices popular, and establish an effective public-private coordination institution suitable for Vietnam’s conditions. AI will not wait, and countries that do not act early will be left behind forever in the education and technology race of the 21st century.

Hoang Anh Tuan

Ambassador, former ASEAN Deputy Secretary-General