Special Report | HKACE Briefing Session on the “AI for Empowering Learning and Teaching” Funding Programme
- wendykwan3
- 2 days ago
- 6 min read
Special Report | HKACE Briefing Session on the “AI for Empowering Learning and Teaching” Funding Programme
Last month, our team had the privilege of attending the briefing session on the “AI for Empowering Learning and Teaching” Funding Programme organized by the Hong Kong Association for Computer Education (HKACE). At the event, we listened to a keynote sharing by Mr. Fong Ting Hei, Council Member of HKACE and Vice Principal of Pope Paul VI College. In this one-hour session, he offered a clear and insightful analysis of the educational vision behind the Government’s HK$500,000 funding programme, and explored how schools in Hong Kong can make good use of AI to drive educational transformation rather than be replaced by it. Below are some of the key takeaways from our on-site coverage.
Core Message: AI as a learning partner, not a threat to teachers
At the start of his talk, Vice Principal Fong pointed out a common misunderstanding in Hong Kong’s education sector: many people worry that AI will “take over teachers’ jobs.” His central argument turned this concern on its head — the true role of AI is to free teachers from the repetitive task of being mere “knowledge transmitters,” so they can evolve into “guides of intelligence.” He stressed that AI should be a learning companion for students, not just a tool for teachers. At present, most schools in Hong Kong use AI mainly to support teachers, such as lesson preparation or automatic marking, but there is still a lack of an ecosystem in which students use AI directly for learning. This, he noted, is exactly where the “AI for Empowering Learning and Teaching” Funding Programme can play a meaningful supporting role.
Teaching, Assessment, Learning
Vice Principal Fong also highlighted the current imbalance in AI education development in Hong Kong, which can be viewed across three dimensions: teaching, assessment, and learning. Among these, teacher-facing tools (such as Gemini、PoE、DeepSeek) are the most mature, assessment applications are developing steadily, while student-facing learning applications remain the least developed — yet arguably the most important.
Teaching: This is the most mature area. Lesson-planning assistants and teaching material generators are already widely used, with a wide range of model choices such as Gemini, DeepSeek, Grok, and PoE.
Assessment: This is the next most developed area. Current applications mainly focus on AI marking systems for language subjects or school-based data analytics systems.
Learning: This is the weakest area, but also the most critical. Students still have limited opportunities to use AI directly, AI literacy development is insufficient, and personalized learning experiences remain underdeveloped.
After outlining this overall landscape, he went on to discuss how schools should choose their technological pathway, including one particularly noteworthy direction: the school-based LLM.
What is a school-based LLM?
During the session, Vice Principal Fong introduced a concept that may be unfamiliar to some, but is highly strategic in significance: the school-based LLM, or a locally deployed large language model. In simple terms, this means running the AI “brain” on the school’s own servers, without depending entirely on external cloud services.Although the concept sounds technical, he explained its three major advantages clearly:
Data stays within the school: This aligns with privacy requirements.
RAG reduces hallucinations: With retrieval-augmented generation, the AI refers to school-provided teaching materials and guidelines before generating answers, making responses more grounded in actual sources rather than fabricated content.
Greater technical autonomy: Schools are not locked into a single vendor and can switch between models such as Llama or DeepSeek more freely.
However, he also acknowledged that a school-based LLM is not the ultimate solution for every school. It requires three important preconditions: at least one technical lead, whether a teacher or IT staff member, who is willing to maintain the system over time; sustained budget for hardware maintenance and teacher training; and an understanding from school leadership that local LLMs may respond more slowly than cloud-based services. If a school does not have these conditions in place, he suggested considering an integrated platform solution instead — a one-stop, ready-to-use option — or subject-specific platforms, or even a hybrid approach combining school-based and cloud platforms.
8 questions for schools to ask vendors
To help schools hold more practical and meaningful discussions with vendors, Vice Principal Fong listed eight key questions that can often reveal a vendor’s maturity and professionalism:
Service model: Is it a one-time purchase or an annual subscription? How will the pricing change after two to three years? Are there any hidden costs?
Student learning integration: Can students use the AI directly, or is it only a teacher tool? How can it be embedded into daily teaching?
Workflow support: Does the platform support customized workflows? This is essential for future AI upgrades.
User interface: Is it easy for both teachers and students to use? Is there a free trial period?
Data privacy: Where is the data stored? Does it comply with Hong Kong regulations? Does the school retain ownership of the data?
Technical support: Does the vendor provide on-site training or only online support? Can teachers operate the system independently?
System integration: Can it connect with the school’s existing LMS? Can it integrate multiple AI services?
Real-world cases: Can the vendor provide authentic examples from Hong Kong schools? What were the actual teaching and learning outcomes
Looking ahead : 2026 and the rise of Agentic AI
Vice Principal Fong gave a summary of the direction AI technology is taking: from the “prompt-based” era in 2024, to the “workflow” era in 2025, and then into the age of “Agentic AI” in 2026. By Agentic AI, he meant systems that can automatically adjust teaching strategies based on students’ past records and proactively track their progress.(Editor's note: The recently popular Clawdbot/OpenClaw is a good example of agent-based AI, demonstrating the speaker's foresight.)
He suggested that schools should prioritize platforms that support workflow design today, while also beginning to accumulate student learning data as a foundation for the next major leap in AI over the coming two years.
Time is tight: A six-month action plan
Vice Principal Fong emphasized that time is already very tight for schools hoping to make full use of this funding opportunity. He recommended that principals take the following steps over the next six months:
February: Form a cross-subject AI task force, identify three initial subjects, and complete the application by the February 28 deadline.
March to April: Gather needs from different subject panels, run trials on multiple vendor solutions, conduct teacher surveys, begin shaping project details, and make an initial budget allocation.
May: Confirm the final technical solution, prepare tender documents, and define evaluation criteria.
June: Receive the funding, launch the tendering process, and arrange teacher training.
July to August: Run a summer pilot, collect feedback, and prepare for full implementation in the new school year.
Closing thought : A Turning Point for Hong Kong's Education
As we left the venue, the line that stayed with us most was Vice Principal Fong's remark: “AI should not only be a tool for teachers; it should be a learning partner for students.” That idea captures the real challenge facing Hong Kong education in the AI era. The Government’s HK$500,000 funding is not meant to encourage schools to buy the latest flashy technology, but to drive a deeper shift in educational practice — from simple knowledge delivery to intelligent guidance.
For school leaders who are still hesitating about whether to invest in AI education, his advice was simple: do not wait for the perfect solution. Start small, pilot early, accumulate data, and prepare for the future. Because the arrival of Agentic AI, much like the internet’s impact on education, is no longer just a possibility — it is becoming inevitable.
Moreover, if the principal and teachers plan to use this “AI for Empowering Learning and Teaching” funding for teaching effectiveness and wants to integrate data, talk to us! The ConnectAR team has ample experiences with AI system development, multi-systems and legacy systems integration, and our leadership has delved into the depths of big database systems and educational systems. For using AI in the Chinese language subject, please can take a look at our ConnectAR AI Brain Lion 🦁. (Free trial available!)
AI Brain Lion 🦁: Your AI assistant for Chinese writing
✅ Especially suitable for personalized teaching assistance in the Chinese subject and enhancing students' AI literacy
✅ 22,000+ intelligent vocabulary database | 300+ idioms with instant analysis
✅ Personalized learning paths | Four-dimensional enhancement of writing skills
✅ Perfectly integrates with AI teaching examples requirements!
Want to know how to incorporate AI Brain Lion into the "Smart Learning" program to support school-based AI applications?
Contact us for inquiries about a free trial, plus other AI integration solutions!
Explore several solutions before making your final choice, and make good use of the funding to enhance teaching and learning effectiveness! 🚀
