以正確的態度、清晰的思維邏輯與 AI 同行 ── 由孩子的未來（Z）倒推今天的學習（A）。 With the right attitude, clear thinking, and human depth — walking with AI from Z (your child's future) back to A (today's learning).
我們不是教 AI ── 我們用 AI，把學校還給老師、把學習還給學生、把孩子還給家長。當機器能代勞越來越多，「人」的價值更需要被守護、被滋養、被看見。 We don't teach AI. We use AI — to give classrooms back to teachers, learning back to students, and children back to families. As machines do more, being human matters more — and must be cherished, nurtured, and seen.
不從 A 開始 ── 從 Z 開始。
從孩子的未來，倒推今天。
We don't start from A.
We start from Z — your child's future — and walk back to today.
傳統教育是 A → Z：跟教科書、跟程序、跟標準答案，由起點按部就班走到終點。但今日的孩子，將要面對的未來，是連大人都未見過的世界。我們必須 從終點開始思考 ── Z 代表孩子十幾年後要成為的人，A 代表今天起步的每一步應如何走。 Traditional education is A → Z: follow the textbook, follow the syllabus, follow the standard answer — walk from start to finish step by step. But the future these children face is one even adults haven't seen. We must start from the end — Z is who the child will become in 15 years, A is how today's first step should be taken.
這套思維方法有正式名稱 ── 教育學界稱之為 「Backward Design」（逆向設計），由 Wiggins 與 McTighe 在 1998 年提出，已被全球頂尖學校採用。我們將這套思維延伸至整個學校的 AI 轉型 ── 由「孩子最終要具備的能力」倒推，再回到今天該如何起步。 This thinking has a name in education: Backward Design — proposed by Wiggins & McTighe in 1998, used by top schools worldwide. We extend this mindset to the entire school AI transformation: start from "what capability the child ultimately needs", then work backwards to what we do today.
AI 將會自動化越來越多的「Z」── 終點、答案、套路。但「A」── 起點、好奇、創造 ── 永遠屬於人。本計劃存在的目的，就是守護孩子心中那個「A」。 AI will automate more and more "Z"s — endings, answers, formulas. But "A" — beginnings, curiosity, creation — will always belong to humans. Our programme helps children hold onto that A.
2026 年 1 月，團結香港基金（OHKF）公佈一份本地調查，受訪人數共 1,200 位 ── 包括 833 位學生與 313 位老師。結果震動教育界。我們不嘗試淡化這些數字，而是選擇正視，並設計出回應方案。 In January 2026, Our Hong Kong Foundation (OHKF) released a 1,200-person Hong Kong survey — 833 students and 313 teachers. The findings shocked educators. We don't try to soften these numbers. We face them, and design our response.
23% 的香港中小學生承認，離開 AI 完全無法完成功課。另有 16% 學生坦承曾將個人資料輸入到 AI 平台。學習的「自主性」正在悄然流失 ── 並非因為 AI 邪惡，而是因為從未有人教導何謂正確使用。 23% of HK primary and secondary students admit they cannot complete assignments without AI. A further 16% admit inputting personal data into AI platforms. Learning autonomy is silently eroding — not because AI is evil, but because no one has taught what proper use looks like.
「Think before you prompt」── 我們課程的第一堂，是教導學生思考三個核心問題：這件事，AI 幫得到我嗎？該不該由 AI 來幫？使用過後，我需要承擔甚麼責任？ "Think before you prompt" — our first lesson teaches students: Can AI help with this? Should it? After I use it, what am I responsible for?
71% 老師擔憂學生過度依賴 AI 會妨礙思考能力。63% 認為 AI 威脅學生的批判思考。只有 3% 老師認為 AI 對學生完全沒有負面影響 ── 換言之，97% 老師都察覺到問題的存在。 71% of teachers worry over-reliance on AI hinders thinking. 63% believe AI threatens critical thinking. Only 3% of teachers say AI has no negative effects — meaning 97% of teachers smell the problem.
我們不主張禁用 AI ── 而是重新設計作業。AI 能輕易解答的題目不出，唯有孩子才能解答的題目才是真正的考核。批判思考並非禁出來的，而是練出來的。 We don't ban AI — we redesign assignments. We don't assign what AI can answer; we assign what only the child can. Critical thinking isn't banned into being — it's practised into being.
95% 學生 + 91% 老師正在使用 AI ── 主要是 Poe、DeepSeek、Copilot、Gemini、Doubao 等公開平台，並非為教育而設計，亦不會特別保護學生資料。OHKF 副研究總監郭凱傑：「光發佈框架是不夠的 ── 必須有持續的跟進。」 95% of students + 91% of teachers are using AI — mostly Poe, DeepSeek, Copilot, Gemini, Doubao — open platforms not designed for education, with no special protection for student data. OHKF Deputy Research Director Victor Kwok: "Publishing a framework alone is not enough — there must be ongoing follow-up."
採用全球公認的「HKU + HKUST 四維框架」── 認識（Know）、運用（Use）、評估（Evaluate）、倫理（Ethics）── 由 HKU Ng/Chu/Qiao 與 HKUST Leung 提出，已被 Queen Mary London、Barnard College 採用。香港研發、世界認可。 We adopt the globally recognised HKU + HKUST four-dimension framework — Know · Use · Evaluate · Ethics — developed by HKU's Ng/Chu/Qiao + HKUST's Leung, now used by Queen Mary London and Barnard College. Hong Kong-born. World-recognised.
市場上充滿販售 AI 工具的供應商。我們不是。我們相信教育的核心永遠是人 ── 是學生的好奇心、是老師的判斷力、是家長的安心、是孩子被看見的需要。AI 只是放大器，不是替代者。 The market is full of people selling AI tools. We are not. We believe the heart of education is — and always will be — people: a student's curiosity, a teacher's judgement, a parent's peace of mind, a child's need to be seen. AI is an amplifier, never a replacement.
AI 給予孩子答案。但答案會過時，懂得學習的人不會。 AI gives children answers. Answers expire. Learners do not.
AI 時代最珍貴的能力，是元認知（Metacognition） ── 知道自己怎樣思考、何時該問、何時該停。我們不教孩子依賴 AI；我們教導他們以 AI 練習思考，學會審視自己的推論、檢查自己的盲點、修正自己的偏見。這是一生受用的能力。 The most precious skill in the AI age is metacognition — knowing how you think, when to ask, when to stop. We don't teach children to depend on AI; we teach them to use AI as a thinking partner, to scrutinise reasoning, spot blind spots, correct bias. A lifelong skill.
未來不屬於懂用 AI 的人，而是屬於懂得何時不用 AI 的人。 The future doesn't belong to those who know AI tools — but to those who know when not to use them.
我們教學生一個判斷框架：這件事，AI 幫得到嗎？該不該由 AI 幫？幫了之後，由誰負責？這套框架可應用於寫作、創作、研究、人際關係、甚至道德兩難 ── 在任何需要創意與靈感的場合。 We teach students a decision framework: Can AI help with this? Should it? Once it does, who is accountable? Applied to writing, creating, researching, relationships, ethics — anywhere creativity and inspiration are needed.
我們教 AI，並非要孩子更像機器 ── 而是要釋放他們去做更像「人」的事。 We teach AI not to make kids more like machines — but to free them to be more human.
同理心、批判思考、溝通、合作、堅韌、好奇心 ── 國際勞工組織（ILO）2025 年 7 月 "Core Skills in the Age of AI" 報告指出：21 世紀技能中，創意、同理心、團隊合作、批判思考在 AI 時代仍然關鍵；同時需要加入新的 AI 素養與倫理理解。技術能力是入場券；軟實力才是天花板。 Empathy, critical thinking, communication, collaboration, resilience, curiosity — the ILO's July 2025 "Core Skills in the Age of AI" report finds that 21st-century skills such as creativity, empathy, teamwork and critical thinking remain critical in the AI age, alongside emerging needs for AI literacy and ethical understanding. Technical skills are the entry ticket; soft skills are the ceiling.
智師學苑並非單一課程，而是一個由六個方向組成的同行計劃 ── 同時為學校、老師、家長、學生四個角色設計。每一個方向均具備實際教學內容、實證研究基礎，以及可量度的成果。 MasterAI Academy is not a single course. It is a companion programme of six pillars — designed simultaneously for schools, teachers, parents, and students. Every pillar has real lessons, research-backed methods, and measurable outcomes.
「Think before you prompt」 ── 教導元認知，而非依賴。 "Think before you prompt" — teach metacognition, not dependence.
以 AI 作為思考夥伴，而非答案機器。學生學會審視自己的推論、檢查自己的盲點、修正自己的偏見 ── 一生受用。 AI as a thinking partner, not an answer machine. Students learn to scrutinise reasoning, spot blind spots, correct bias. A lifelong skill.
何時 ・ 為何 ・ 如何 ── 一套帶得走的判斷框架。 When · Why · How — a portable decision framework.
應用於寫作、創作、研究、人際關係、道德兩難 ── 在任何需要創意與靈感的場合。 For writing, creating, researching, relationships, ethics — anywhere creativity and inspiration are needed.
上堂前學知識 ・ 上堂時學思考。 Knowledge before class · thinking during class.
用 AI 為老師備課、為學生預習。把課堂時間還給討論、實作、思辨 ── 提升學生投入度，自然提升收生競爭力。 AI helps teachers prep and students preview. Class time goes to discussion, hands-on, debate — engagement up, enrolment competitiveness up.
同步 ・ 同頻 ・ 同一頁。 Same pace · same page · same channel.
為家長設計的簡短同步課程，與孩子的學習進度緊密配合 ── 在飯桌上能對話，在書房中能輔助，於深夜時能安心。 A short parent companion track that runs alongside the children's curriculum — talk at the dinner table, help in the study, sleep at night.
並非培訓老師使用 AI ── 而是以 AI 培訓老師。 Not training teachers to use AI — using AI to train teachers.
桑德蘭大學認證的進階課程，配合 AIPPA 導師網絡實戰陪伴。撥款結束之後，能力與資歷依然屬於老師、屬於學校。 Sunderland-accredited certificate paired with AIPPA mentor network. After the grant ends, the credentials stay with the teacher and the school.
AI 取代不了的，才是孩子的未來。 What AI can't replace — is your child's future.
同理心、批判思考、溝通、合作、堅韌、好奇心 ── 並非另設一個課程，而是融入每一節 AI 課的靈魂。 Empathy, critical thinking, communication, collaboration, resilience, curiosity — not a separate course, but the soul woven into every AI lesson.
我們的方法並非一次性培訓，而是五步階梯式設計 ── 每一步均設有微任務、Mentor 回饋、成果驗收。學生、老師、家長同步遵循這五個步驟，因此「同頻」是真正的同頻。 Our method is not a one-off training. It's a five-step ladder — every step has micro-tasks, mentor feedback, tangible output. Students, teachers, and parents follow the same five steps. That's why "same page" actually means same page.
由零開始，認識 AI 是甚麼、能做甚麼、不能做甚麼。在生活情境中辨認 AI 的影子。From zero. What AI is, what it can do, what it can't. Spot AI in everyday life.
學會提問、下指令、拆解任務 ── 並非單純 prompt engineering，而是思考本身的訓練。Ask. Instruct. Decompose. Not just prompt engineering — training the mind.
每日小練習，將 AI 由概念變成日常工具。撰寫一封電郵、整理一份筆記、解開一道題目。Daily small tasks. AI from concept to everyday tool. Write an email, organise notes, solve a problem.
應用於真實的學習、創作、溝通場景，產生有意義的成果。可以是一份報告、一場演講、一份作品集。Apply to real learning, creative, communicative settings. A report. A speech. A portfolio.
AIPPA 義工 Mentor 陪伴、回饋、解答疑難。學習不再孤單，能力得以持續。AIPPA volunteer mentors provide companionship, feedback, answers. Learning is no longer alone. Capability endures.
這並非另一份「列出 21 世紀技能」的清單。我們將學界共識的軟實力，落地分成三個角色 ── 學生、老師、家長 ── 各自需要建立的核心能力，再嵌入每一節 AI 課中。不另設科目，不加重負擔。 This isn't another "21st-century skills" list. We took established research, broke it into the three roles — student, teacher, parent — each with their own core skills, and wove them into every AI lesson. No new subject. No extra load.
當 AI 代勞一切時 ── 是甚麼，讓你成為你？When AI does the doing — what makes you, you?
不僅是教 AI，更是以 AI 重新成為老師。Not teaching AI — using AI to be a teacher again.
無需精通 AI ── 但須懂得聆聽。You don't need to know AI — you need to listen.
我們不會以貴校作為試驗田。每一個方法、每一份教材，均已在真實環境中實踐、修正，並由學生與老師親身驗證。 We don't experiment on your school. Every method, every material, has run through real classrooms — refined, tested, validated by real students and teachers.
起點。將 AI 教育帶入新興國家的大學課堂。經驗證明：在跨文化、跨語言、基建有限的條件下，五步學習法仍然有效。 The beginning. Brought AI education to a developing-nation university. Proved: cross-cultural, cross-language, low-infrastructure — our five-step method still works.
非洲基層社區 6 日深度集訓，並於同年進行英國教育交流。此等經驗直接應用於後來香港的少數族裔社群與基層學生支援工作。 6-day grassroots intensive in Africa + UK education exchange. Lessons later applied directly to Hong Kong's ethnic-minority communities and grassroots students.
首 4 個月 20+ 場，覆蓋國際學界、本地企業、專業公會、社福機構、基層社群五大範疇。月均 6–7 場。 20+ in the first 4 months, across international academia, local enterprise, professional bodies, NGOs, grassroots. Average 6–7 per month.
市場上大多數供應商，是為學校而生的 EdTech 公司。我們不是。ZtoA 是一間已為 ICAC、AXA、騰訊、P&G、Microsoft、Hitachi 等機構服務超過 20 年的 AI 與商業諮詢公司。教育，是我們將企業執行紀律延伸至社會的方向 ── 並非我們的副業。 Most providers in this market are EdTech-for-schools companies. We are not. ZtoA has been an AI and business consultancy for 20+ years — serving ICAC, AXA, Tencent, P&G, Microsoft, Hitachi and others. Education is where we extend our enterprise execution discipline to society — not our sideline.
為政府、銀行、跨國企業執行 AI 與品牌轉型的紀律：完整文件、可量度成果、合規可審計。同樣的紀律帶入學校。The same discipline we apply to government, banks, multinationals — complete documentation, measurable outcomes, audit-ready compliance — brought to schools.
雖然不能以撥款支付 ── 但其權威性無可置疑。教師完成 ZtoA × 英國桑德蘭大學（香港）× ACMCP 認證後，獲得國際認可的學歷資歷。撥款結束後，能力與資歷依然留存。Cannot be paid from the funding — but carries international authority. Teachers earn a recognised credential through ZtoA × University of Sunderland HK × ACMCP. The funding ends; the credential remains.
並非單純合作關係 ── 這是我們親自創建的協會。專業導師義工網絡可直接調動，公開課、分享會的聽眾與媒體曝光得以一站到位。Not a partnership — an association we built. Professional mentor volunteer network at our direct disposal. Audience and media for open classes and sharing sessions in one place.
我們的 AI 諮詢業務專注於 HKMA、SFC、IA、PCPD、DPO 的監管要求。同樣的嚴謹度應用於學校 AI 使用政策、家長同意書、學生資料保護等領域。Our AI consulting practice specialises in HKMA, SFC, IA, PCPD, DPO compliance. The same rigour applies to school AI policy, parent consent, student data protection.
我們不依靠售賣特定平台賺取分成。工具建議完全因應貴校的實際情況而定，包括預留與未來教育局可能推出之中央平台的兼容空間。We earn no platform commission. Tool recommendations are purely based on what fits your school — including future-proofing for any central platform EDB may launch.
30 分鐘對話，沒有銷售壓力。請告訴我們貴校的真實情況，我們則分享在 200+ 個機構觀察過的盲點與機會。即使最終您決定不與我們合作，我們仍會留下一份「校本 AI 思考清單」予貴校 ── 它本身已足夠有用。 A 30-minute conversation. No sales pressure. Tell us your school's real situation; we'll share blind spots and opportunities from 200+ organisations. If you decide not to work with us, that's fine — we'll still leave you a "School AI Thinking Checklist". It's useful on its own.