From the ward to the screen
How it started
I'm an infectious disease physician. Antimicrobial resistance isn't a headline for me — it's something I deal with every day.
Honestly, I never planned to make a short film. Then I came across a competition organized by Taiwan's Ministry of Health — an AI animation contest for healthcare professionals. Antimicrobial resistance immediately felt like the right subject, because this is something the public needs to see, not just doctors. But how do you make people without a medical background feel the pressure? Articles are too dry. Health education videos are too boring. So I decided to let the visuals speak.
這部片怎麼來的
我是感染科醫師，抗藥性對我來說不是新聞標題，而是每天都在面對的日常。
老實說，一開始根本沒想過要做短片。某天看到衛福部舉辦了一場醫護同仁 AI 動畫繪圖競賽，腦中馬上浮現抗藥性這個議題——「這件事，不應該只有醫療人員知道。」但問題是：怎麼讓沒有醫學背景的人也能感受到這份壓力？寫文章太硬，做衛教太無聊。那就用畫面來說吧。
Why cinema, not education?
There's a gap between knowing something and feeling it. I could tell you that antibiotic resistance kills over a million people a year. But numbers don't make you pause. A face does. A moment of hesitation does.
That's why I chose the language of film — not to teach, but to create a space where you stop scrolling for a moment.
為什麼不做衛教，而是做短片？
數據誰都會講，但不會有人因為一個數字就停下來思考。
我真正想做到的是：看完這部片之後，下次再聽到「抗藥性」三個字，你腦中浮現的不再是圖表——而是一張臉，和一隻細菌怪物。
The AI part
Let me be direct: I don't know how to make films. I can't do 3D. I can't afford a crew. AI made this project possible — MidJourney for the visuals, Kling AI for the motion.
But here's what people often misunderstand: AI generates fast, but choosing is slow. Just for the main character's look and the bacteria designs, I went through countless variations in MidJourney before I could decide. AI doesn't know what's right — it only knows what's possible. I used many generated images to create video clips, but only a handful actually made it into the final cut.
關於 AI 這件事
說實話，我不懂拍電影，不會 3D，也請不起劇組。是 AI 讓這一切變得可能。
但有一件事很多人容易誤解：AI 生成很快，選擇卻很慢。光是主角的形象和細菌的造型，我在 MidJourney 就做了非常多張才拍板定案。AI 不知道什麼是「對的」，它只知道什麼是「可能的」。最後拿去生成影片的圖有很多張，但真正剪進短片裡的，只有少數幾個鏡頭。
The visual choices
I made the bacteria look monstrous on purpose. In real life, you can't see them — that's the whole problem. So I turned them into something you can't not see.
The microscopic battlefield — white blood cells versus pathogens, drugs deployed like weapons — none of this is science fiction. It's happening inside patients' bodies every single day. I just scaled it up until you couldn't ignore it.
Antibiotic misuse, improper prescriptions, patients stopping treatment early — these are how bacteria learn to survive. I used tight pacing and shifts between warm and cold tones to convey this. Pharmacies, fish farms, livestock operations — these scenes exist all around us.
In the end, there's still hope. New drugs are being developed — we still have weapons. But the threat of resistance lurks in every corner.
視覺上的選擇
我故意把細菌畫得像怪物。因為在現實中你根本看不到它們——而這，才是最可怕的地方。
影片裡的微觀戰場——白血球與病原體正面交鋒、藥物如同武器被投入前線——這些不是科幻想像，而是每天在病人體內真實上演的事。 我只是把它放大到你沒辦法假裝沒看到。
抗生素濫用、不當使用、中途停藥——這些行為正在教會細菌如何活下來。我用緊湊的剪輯節奏和冷暖色調的交替來傳達這份急迫感。藥局、魚塭、畜牧場——這些場景就在我們的生活裡，一點都不遙遠。
但最後，我們還是要留住希望。新藥研發成功，代表手中還有武器。只是抗藥性的威脅，從未真正離開。
What I learned
Making this film reminded me of something I already knew as a doctor but sometimes forget: tools don't create meaning. Attention does.
AI gave me images I couldn't make by hand. But the story came from standing at a patient's bedside, wondering if the next antibiotic would work.
做完這部片之後
這部片讓我重新想了一件事：重點從來不是工具有多厲害，而是你把目光放在哪裡。
AI 給了我以前做不到的畫面，但故事的核心，來自站在病床邊，不確定下一支抗生素還有沒有效的那個瞬間。
Tools & Process | 創作工具
- Video: Kling AI, MidJourney
- Image & Style: MidJourney
- Editing: CapCut
— TingYu's AI Art · Superbug Crisis · 2025.10 · Taiwan
