In July 2025 the General Office of the Ministry of Education issued the Notice on Organising the Implementation of the Digital-Empowerment Teacher Development Action (Teacher Hall Letter [2025] No. 13). It sets out to expand the supply of quality resources and services, driven by the integrated and innovative use of digital and artificial intelligence technologies, and to open new tracks for teacher development. This is not a document about what equipment to buy; it is a document about who uses AI, how they use it, and where the boundaries lie.

From “connected devices” to “teacher capability”

In the earlier phase of education digitalisation the focus was on infrastructure and platforms. This notice shifts the focus to people: whether a teacher can understand the capabilities and limits of AI tools, and can judge which steps may be handed to a model and which must stay with a human, becomes the new standard of capability.

That changes what schools must treat as their compliance object. In the past, compliance in education informatisation revolved around cybersecurity, data classification and the protection of minors; now it also includes norms for how teachers use AI — in lesson preparation, question-setting, marking and learning analytics.

Three footholds for AI in the teaching process

Content. When a teacher produces courseware, exam questions or comments with generative tools and then publishes or uses them for teaching, content-labelling requirements may be triggered. See our notes on the Measures for Labelling AI-Generated Synthetic Content and the basic safety requirements for generative AI services.

Data. Learning analytics and automated marking usually involve students’ personal information, much of it belonging to minors. The selection and use of teacher-side tools should be assessed against the standards set out in our note on the obligations of major personal information processors.

Ethics. Algorithmic evaluation and emotional-companionship applications used in education should be assessed against our note on the AI education ethics reference framework.

Teachers are not bystanders to the tool

It is worth stressing that teachers are not passive users in this round. When a teacher brings AI-generated content into the classroom, that teacher bears professional responsibility for its authenticity, accuracy and appropriateness. Technology can share the workload, but it cannot transfer the responsibility.

This view is consistent with recent judicial and regulatory positions: technological neutrality does not mean application neutrality, and the party who applies a tool is responsible for the consequences in the specific setting. See our note on the Hangzhou case on unfair competition in AI commercial application.

Where schools and education institutions can start

First, build a scenario-by-scenario list for teacher use of AI, clarifying which steps may be assisted and which must not be automated, such as the final conclusion of an academic assessment.

Second, merge digital-literacy training with compliance training, so that “knowing how to use” and “using compliantly” do not drift apart.

Third, map the flows of student data, especially cross-border transfers and entrusted-processing arrangements when third-party AI education products are introduced.

What makes the education setting special is that it is governed by three sets of rules at once — content regulation, personal information protection and minors’ protection. How fast AI enters the classroom ultimately depends on whether those three sets of rules can be translated into enforceable norms of teacher conduct. See our note on China’s internet information service measures for intelligent services.

This article is based on public information on the Ministry of Education General Office’s Digital-Empowerment Teacher Development Action notice and is for reference only; it does not constitute legal advice.