Vietnam Classifies Some AI in Education as ‘High Risk’

Vietnam has placed several uses of artificial intelligence in education on its new list of high-risk AI systems.

The July 2026 rules include AI systems that automatically test, evaluate or rank learners, systems that monitor and analyse learner behaviour, and some self-study tools that generate educational content from uncontrolled data sources.

A classroom in Vietnam. The country’s new AI framework identifies several education applications as high risk, particularly where systems make or influence judgements about learners. Image: Phat14082005 / Wikimedia Commons, CC BY-SA 4.0.

High risk does not mean banned

The classification matters because it separates relatively low-stakes classroom experimentation from uses that can affect a student’s opportunities, record or treatment. The rules do not amount to a blanket ban on educational AI, but they signal that some applications deserve stronger safeguards and oversight.

For international schools, the distinction is useful even outside Vietnam. An AI tool that helps a teacher generate discussion questions carries a very different level of risk from one that scores students, ranks them or continually analyses their behaviour. Schools need governance that reflects that difference.

Read the Vietnamese Government announcement