Executive Summary
Artificial intelligence (AI) provides an unprecedented technical foundation for the long-cherished educational ideal of “teaching in accordance with each student’s aptitude” (因材施教). As the government rolled out the “AI-Enabled Learning and Teaching” (「智啟學教」) funding programme, published the Blueprint, and disbursed a one-off grant of HK$500,000 to each publicly-funded school between 2025 and 2026, Hong Kong’s AI-in-education agenda has shifted from a question of “whether to use AI” to one of “how to use it well.” To examine the institutional challenges schools face at this turning point, the research team visited more than 80 primary and secondary schools during the same period and analysed over 200,000 de-identified pieces of student work produced with AI.
The study finds that Hong Kong schools typically operate multiple AI platforms in parallel, with no common data formats, standards, or governance mechanisms linking them. This has produced large numbers of disconnected “data silos” (數據孤島), making it difficult for teachers, parents, and schools themselves to gain a complete picture of students’ learning journeys or to assess objectively whether AI is actually improving learning. What Hong Kong’s education sector truly lacks is not AI, nor even learning data, but an institutional framework that allows learning data to accumulate continuously, circulate safely, and support teaching interaction. The study therefore puts forward three short-to-medium-term recommendations: (1) schools should establish school-based directions for AI and data governance; (2) school sponsoring bodies should set unified data-governance requirements to reduce the duplicative cost of each school building its own regime from scratch; and (3) the education sector, universities, and professional bodies should jointly establish a voluntary evaluation and accreditation mechanism for AI solutions, gradually cultivating a shared culture of using AI well, so that AI can be transformed into the educational infrastructure that genuinely supports “teaching in accordance with each student’s aptitude.”
1. Background
In recent years, Hong Kong has actively promoted AI in education. The government earmarked HK$2 billion in 2025 to drive digital education and, in December 2025, launched the “AI-Enabled Learning and Teaching” funding programme. On 17 June 2026, the Education Bureau published the Blueprint and, on the same day, disbursed a one-off grant of HK$500,000 to each publicly-funded school. Hong Kong’s AI-in-education agenda is moving from a question of “whether to use AI” to one of “how to use it well.” However, after spending a year visiting more than 80 primary and secondary schools and analysing over 200,000 de-identified pieces of student work produced with AI, our research team found that the key factor determining a school’s effectiveness in learning, teaching, and assessment is not how advanced the AI tools it adopts are, but whether the learning data generated by AI can be preserved, integrated, and used on a continuous basis.
2. Findings
The study finds that Hong Kong schools typically operate multiple AI platforms in parallel, and that the lack of common data formats, standards, and governance mechanisms across these systems has caused students’ learning records to become extremely fragmented, giving rise to large numbers of disconnected “data silos”. Teachers struggle to gain a complete picture of students’ learning journeys; parents find it hard to obtain a comprehensive view of their children’s development; and schools themselves find it difficult to assess objectively whether AI is truly improving learning outcomes. In other words, what Hong Kong’s education sector truly lacks is not AI, nor learning data, but an institutional framework that allows learning data to accumulate continuously, circulate safely, and support teaching interaction.
3. Policy Recommendations
Drawing on the experience of different countries, we believe Hong Kong must address the problems caused by “data silos” in education without delay, and should make full use of the strengths of its education system and its free market to build, step by step, a governance regime for AI learning data. The study puts forward three concrete, short-to-medium-term recommendations:
- Schools should take the lead by drawing on the “Frequently Asked Questions on Internal Data Governance” to establish school-based directions for AI and data use.
- School sponsoring bodies should set standardised data-governance requirements, reducing the cost of each school having to build its own system from scratch.
- The education sector, universities, and professional bodies should jointly establish a voluntary evaluation and accreditation mechanism for AI solutions, gradually cultivating a shared Hong Kong education-sector culture of using AI effectively and sustainably.
AI should not be regarded as just another piece of educational technology, but should become a new foundation that supports teachers in practising “teaching in accordance with each student’s aptitude”. Under “One Country, Two Systems” Hong Kong stands at an important turning point in the development of AI in education. Building a culture of using AI effectively and sustainably today will inevitably form part of tomorrow’s educational infrastructure. Only when students’ learning journeys can accumulate continuously, circulate safely, and be deployed effectively will the data value generated by AI be truly translated into educational value, propelling Hong Kong’s education from “digitisation” to genuine “intelligence”.
(Translation supported by AI)
