公益醫療知識平台・資訊經專業編輯整理

關於我們

繁中

/

EN

人工智慧(AI)會取代醫師嗎?

人工智慧正在快速重塑醫療與醫學教育,從診斷、住院醫師訓練、人力配置到外科手術,AI 的角色日益重要。然而,科技真正改變的或許不是「醫師是否會消失」,而是「未來需要什麼樣的醫師」。AI 不會取代醫師,但善用 AI 的醫師,可能取代拒絕擁抱科技的醫師。

自從 ChatGPT 於2022年公開問世以來,年輕人規劃職涯的方式便發生了劇烈的轉變。

數個世代以來,我們總是告訴下一代:求學職業的選擇,應該追隨自己的熱情,相信自己的內心。 然而,在這個前所未有、變化速度極其迅猛的時代——從 Gemini、Claude 的出現,到為各行各業量身打造的專業人工智慧工具——許多傳統職業正面臨前所未有的生存危機。作家、律師、設計師與會計師,都站在自動化取代人力的第一線。

與此同時,隨著全球人口以前所未見的規模快速老化,基層醫療與家庭醫學對醫師的需求也持續飆升。2016年,the Nobel prize winner in 2024Geoffrey Hinton 曾預測,人工智慧將在五年內使放射科醫師失去存在的必要性。事後看來,這個時間表顯然過於樂觀。然而,如果因此認為醫師將永遠無可取代,同樣是一種天真的想法。
我們只需看看今日街頭,無人駕駛汽車正以極快的速度逐漸成為現實,便能理解:未來往往比我們想像得更快抵達。

歷史告訴我們:科技進步必然重新塑造人類的勞動方式

歷史一再提醒我們,每一次重大的科技躍進,都會重新定義人類工作的樣貌。1950年代的曼哈頓,高樓大廈裡仍由專職的電梯操作員負責控制電梯。當自動電梯最初進入市場時,一些心存恐懼的住戶甚至威脅,如果電梯裡不再有人工操作員,他們就要搬離大樓。

大約在同一時期,每一家大型醫院、企業與公司,都高度依賴電話交換機接線員來協助轉接電話。

工業革命如此;此後的每一次數位革命亦然——科技不斷改變就業市場。但是,我們並沒有因此消失;我們學會了適應,才能繼續生存。

如今,AI 正跨越門檻,進入醫療的每一個角落

人工智慧正在跨越一道又一道門檻,深入醫療體系的每一個環節,甚至從醫學教育的最初入口——醫學院招生——便已開始發揮作用。

2018年,紐約大學(NYU)宣布提供免學費的醫學教育後,申請人數迅速攀升至數萬人之多,每一個名額甚至面對超過百人的競爭。

面對如此龐大的申請量,若要同時兼顧公平性與效率,人工智慧驅動的篩選系統便成為不可或缺的工具。

而在招生之外,AI 更正在從根本上重新塑造臨床醫療的運作方式:

診斷精準度(Diagnostic Precision)

在病理學等專業領域中,醫師必須透過顯微鏡檢視組織切片,進而做出攸關病人生命的重要診斷。AI 能夠大幅縮短判讀與報告所需的時間,同時顯著提升工作效率與診斷的精確度。

住院醫師訓練最佳化(Residency Optimization)

智慧排班模型可以最佳化住院醫師的值班時數,並確保受訓醫師能夠均衡接觸各種複雜且具有高度學習價值的病例。

當住院醫師遇到罕見的臨床情境時,AI 也能成為即時的知識與參考引擎,協助其進行診斷推理。

醫療人力配置(Workforce Distribution)

AI 分析工具能夠量化不同醫療專科的人力需求,並將醫師人力更精準地與社會實際需求相互配對,進而促進更廣泛的醫療公平與資源均衡。

更重要的是,AI 能填補我們的不足,卻無法取代我們的人性。

演算法模型目前仍存在許多盲點、演算法偏見,以及所謂的「幻覺」(hallucinations)——也就是AI產生看似合理、實則錯誤或虛構資訊的情況。因此,人類的思維與情感依然不可或缺。最終的醫療決策,需要人類去修正、理解脈絡、權衡風險,並最終做出判斷。

外科訓練正是這場轉型最明顯的縮影

這場變革,在外科醫學訓練領域尤其明顯。

傳統醫學教育中經典的 「看一次、做一次、教一次」(see one, do one, teach one) 模式,如今已不再實際,也不再是唯一必要的訓練方式。

現代醫學訓練者可以在真正替人類病人進行第一次切開之前,就透過沉浸式的3D模擬平台,在AI監督與輔助下完成訓練,而不必再完全依賴傳統的動物解剖訓練。

科技如何改變外科醫師的養成

回顧我自己的外科生涯,從傳統手術逐步轉向微創腹腔鏡手術(laparoscopy)、血管內動脈瘤修復術(EVAR),以及 Da Vinci 達文西機器人手術,每一次技術轉型,都需要投入大量而高度專業化的訓練。

這些重要的里程碑,使我深刻體會到:先進科技能夠如何提升病人的治療成果。

而如今,透過整合飛行模擬器所採用的訓練方法,以及結合影像技術的手術虛實整合環境,AI 正在大幅縮短外科醫師從受訓者走向成熟臨床醫師所需的時間。我早年接受訓練時,常聽到一句老話:「日子過得很長,但歲月過得很快。」今天,這句話已不再以過去那種令人筋疲力竭的方式適用於外科訓練。
那麼,最終的問題來了:AI 會取代主刀醫師嗎?

對於例行性、低複雜度的手術而言,自動化執行的時代,可能比許多人想像得更快到來。然而,當面對高風險、高複雜度的醫療情境,需要細膩的臨床判斷、複雜的決策能力,以及真正的人性同理心時,機器仍無法取代外科醫師的雙手與心。

簡而言之

AI 不會取代醫師。但善用 AI 的醫師,終將取代那些拒絕使用 AI 的醫師。

真正通往未來的道路,並不是人類與人工智慧彼此競爭,而是學會如何掌握兩者共存的能力。未來屬於那些既懂得行醫,也懂得駕馭 AI 的人。

Will AI Replace Doctors?

Ever since the public debut of ChatGPT in 2022, the way young people plan their careers has shifted dramatically. For years, we advised the next generation to follow their passions and trust their hearts. Yet, in an era of unprecedented rapid change—marked by the arrival of Gemini, Claude, and specialized tools tailored for every industry—many traditional professions face existential uncertainty. Writers, lawyers, designers, and accountants find themselves on the front lines of automated displacement.

Meanwhile, as global populations age at a scale never seen before, the demand for primary care physicians continues to skyrocket. When the Nobel winner in physic of 2024, Geoffrey Hinton predicted in 2016 that AI would render radiologists obsolete within five years, his timeline was grossly exaggerated.

However, it would be equally naive to believe that doctors remain entirely irreplaceable. We need only look at our streets today, rapidly filling with autonomous vehicles, to see how fast the future arrives.

History reminds us that technological leaps inevitably reshape human labor. In the 1950s, Manhattan high-rises were manned by elevator operators. When automated elevators first entered the market, fearful residents threatened to move out if the human operator was removed. Around the same period, every major hospital, business, and corporation relied strictly on telephone switchboard operators to connect calls. The Industrial Revolution—and every digital revolution since—altered the job market. We did not disappear; we adapted to survive.

Today, AI is crossing the threshold into every corner of medicine, beginning with the very entry point starting from medical school admissions. When NYU announced tuition-free medical education in 2018, applications surged into the tens of thousands—yielding over a hundred applicants for every single slot. Managing such volume fairly and efficiently became possible only through AI-driven screening systems.

Beyond admissions, clinical practice is being fundamentally reshaped too. Diagnostic Precision: In fields of pathology, where specialists examine tissue slides under microscopes to render critical diagnoses, AI accelerates reporting time while dramatically improving efficiency and precision.Residency training Optimization: Intelligent scheduling models optimize on-call hours and ensure trainees get balanced exposure to complex, high-yield cases. When a resident encounters a rare clinical scenario, AI acts as an instant reference engine to aid diagnostic reasoning.

Workforce Distribution: AI analytics can quantify specialty demands and match physician manpower directly to societal needs, promoting broader healthcare equity.

Crucially, AI fills our gaps, but it does not replace our humanity. While algorithmic models still suffer from blind spots, algorithmic bias, and hallucinations, the human mind and heart remain essential to refine, contextualize, and perfect the ultimate decision.

This transformation is nowhere more evident than in surgical training. The classic medical paradigm of “see one, do one, teach one” is no longer practical or necessary. Trainees today bypass traditional animal dissections in favor of immersive 3D simulation platforms supervised by AI before ever making an incision on a human patient.

Reflecting on my own surgical career, transitioning into minimally invasive laparoscopy, endovascular aneurysm repair (EVAR), and Da Vinci robotic surgery required intensive specialized training. These milestones offered me deep insights into how advanced technology elevates patient outcomes. Now, by integrating flight-simulator methodologies and video-augmented surgical reality, AI is drastically shortening the timeline to surgeon maturity. The old adage from my early training days—that “the days are long, but the years are short”—no longer applies in the same grueling way.

This brings us to the ultimate question: Will AI replace the primary surgeon? For routine, non-complicated procedures, automated execution is approaching faster than many realize. But in high-stakes scenarios demanding nuanced clinical judgment, complex decision-making, and genuine human empathy, a machine cannot substitute for a surgeon’s hands and heart.

In short, AI will not replace doctors—but doctors who utilize AI will inevitably replace those who do not. The true path forward lies in mastering this coexistence.