Go
Deep
Deep
[16:32] The students tackled two motions they had never prepared for, and still argued them coherently - the real point of debate training is not memorising a rehearsed position, it is building the ability to think deeper on the spot, which is the underlying question education itself should be asking.
[16:31] Counter view: visual art should be a platform for creative expression, and using AI to express wild, imaginative ideas can be part of that - though it may run against the specific hands-on technique visual art is meant to teach.
[16:30] A hands-up experiment: many who never scored top marks in visual art, and were told their drawing was ugly, still call drawing their hobby - a reminder that Hong Kong's art classes may train technical skill without necessarily building a love of it, and art education should aim to make everyone enjoy art, not knock their confidence.
[16:29] One view: AI can spark early ideas, but drawing itself comes from lived, personal experience that AI cannot feel - AI creation ultimately stays a text-based exchange, while visual art creation is fundamentally different, so traditional visual art should not lean on AI.
[16:28] When a coach asked students to prompt an AI image generator for "precious friendship," one output literally rendered the words "precious friendship" as text on a blackboard rather than depicting the concept - not really an AI failure so much as a mismatch: traditional art training teaches technique and composition, while AI image generation only takes a text prompt.
[16:27] Core question raised: should visual art class stay pen-and-paper only, or embrace AI - and more importantly, how many students who are genuinely creative and have a good eye but are not skilled at manual technique feel discouraged (getting a C, no sense of achievement) rather than falling in love with art? Could AI remove a barrier rather than replace the skill?
[16:27] Humanoid AI that talks and looks human already exists, and physical gestures like hugging may not be far off - meaning whoever comforts you in the future might not be human at all, but could feel like it. Debate is not just about arguing a position, it is a way to force yourself to confront a scenario before it actually happens.
[16:26] One view: many people today lack real companionship and emotional support, so if paying for AI company genuinely lifts someone's mood and helps them through a dark patch, that should be permitted - AI playing a companion role to help someone feel better is different from actively encouraging reliance on it.
[16:25] Paid AI companionship still has a pushback view: real friendship needs real human interaction, and AI can only offer cold text, not the human warmth of something like a comforting hug after a bad exam.
[16:24] Students raised a catch nobody expected: AI companionship ultimately runs on a company's paid subscription model, so if a user genuinely feels emotionally dependent on an AI companion, they are paying ongoing money to sustain that relationship - is that acceptable, especially for younger users?
[16:22] The debate format does not lock students into one fixed side - preparing both angles means genuinely understanding why the other side supports or opposes a motion, and a debater's own view can shift partway through.
[16:22] The debate topic has shifted from "should teenagers date" to "can humans date AI" - students separate three distinct questions that easily get conflated: should it be permitted, should it be encouraged, and will it become mainstream. Permitting something is not the same as encouraging it or pushing it toward the mainstream.
[16:21] Despite AI's huge data base, it still leans on surface-level text understanding rather than truly feeling a patient's pain - human doctors' edge is reading facial expressions and body language during consultation and probing further into a patient's daily life, a humanised, back-and-forth diagnosis AI cannot yet replicate.
[16:20] Nurses already play a role similar to first-line triage, asking screening questions - once AI can understand a patient's meaning almost without error, handing that first triage step to AI starts to feel acceptable to some.
[16:19] On AI doing preliminary diagnosis in emergency rooms: Cantonese patients describe pain with highly local, subjective terms (like "chek chek tung", "yan yan tung"), and it is unclear whether AI can genuinely understand these or whether that reading should stay with human doctors.
[16:19] Personal AI-chat time can quietly edge out time with friends and family - many people do not notice how much their AI-search time has grown, and managing that trade-off is a real test given how little free time people in Hong Kong already have.
[16:17] Students increasingly treat AI as a teacher rather than just a tool - after finishing homework they ask AI to check right or wrong and explain the mistake, the same role IB top scorers described using AI for.
[16:16] Traditional teaching still trains students to compute answers by hand, but when a student asks AI directly the moment a question comes up, it exposes how fast younger students already reach for AI instinctively - schools still have not caught up to how naturally this generation collaborates with it.
[16:15] Students will not compete against AI after graduation, they will work alongside it - critical thinking training (like this debate competition) breaks a complex question into its hidden sub-questions rather than jumping to yes/no, the same skill AI adoption and design thinking both need.
[16:12] AI can gather and organise information, but judgment, filtering and decision-making stay with the student - the future skillset is asking good questions, actively questioning information, and building original ideas on top of what AI produces.
[16:09] Students can no longer choose whether to use AI - many primary and secondary students are already using it well before the debate about "should we allow it" catches up.
