Name: 除了會生成吉卜力圖像，AI 還能造福我們的「健康」嗎？ (How Is AI Revolutionizing Medicine?)
Uploaded: 2025-04-25T21:01:20.000Z
Duration: 7 min 51 s

When we talk about AI in healthcare, I think we need to understand it as a foundational change in the toolkit we have available.

當我們談論醫療保健領域的人工智慧時，我認為我們需要將其理解為我們現有工具包中的一項基礎性變革。

情態副詞

In the same way, and I mean this profoundly, that algebra changed our understanding of the world and of math.

同樣，代數改變了我們對世界和數學的理解，我的意思很深刻。

In our field, the major changes are happening in diagnostic testing.

在我們的領域，診斷測試正在發生重大變化。

How we can now interpret them more effectively and without as much expertise needed.

我們現在可以如何更有效地解讀它們，而不需要那麼多的專業知識。

We are seeing the old patterns of care change into new ways of doing care.

我們看到舊的護理模式正在轉變為新的護理方式。

The topic of how is AI going to be used is so interesting because these new machine learning, sophisticated models, they aren't the end of the question, it's the beginning of the question.

如何使用人工智慧這個話題非常有趣，因為這些新的機器學習、複雜的模型並不是問題的終點，而是問題的起點。

AI is the future of healthcare, or so we've been told.

人工智慧是醫療保健的未來，我們一直被這樣告知。

And how does it have the potential to affect you?

Artificial intelligence has, in some form, been around for decades.

人工智慧已經以某種形式存在了幾十年。

But most people are just finding out about it because of things like ChatGPT, which have come into the mainstream.

但大多數人都是通過 ChatGPT 等方式才瞭解到這一點的，而 ChatGPT 已經成為主流。

In healthcare, artificial intelligence has been around for years as well.

在醫療保健領域，人工智慧也已存在多年。

It's helped doctors transcribe notes faster.

It's helped with patient experiences with chatbots.

聊天機器人為患者體驗提供了幫助。

But with generative AI, it's really supercharging what healthcare professionals and scientists are able to do by helping predict future health risks, make diagnoses faster and more accurately, and also helping with potential drug discovery for new novel treatments and therapeutics.

但是，有了生成式 AI，醫療保健專業人員和科學家就能通過幫助預測未來的健康風險、更快更準確地做出診斷，以及幫助潛在的藥物發現新的新型治療方法和療法，從而真正提高工作效率。

The AI boom really hit healthcare in the last five years.

近五年來，人工智慧熱潮真正席捲了醫療行業。

We've actually been able to transition into having the models or the tools we built to actually work with the data very effectively and have been able to define new ways to treat diseases, new ways to diagnose diseases, and finally find individualized signatures of how I as a patient or somebody else as a patient would do if they got a certain treatment or therapy.

實際上，我們已經能夠過渡到使用我們建立的模型或工具來非常有效地處理數據，並且已經能夠定義治療疾病的新方法、診斷疾病的新方法，最終找到個性化的特徵，即我作為病人或其他人作為病人，如果接受某種治療或療法，他們會有什麼表現。

The gen AI boom is assisting doctors in more advanced ways.

人工智慧熱潮正以更先進的方式為醫生提供幫助。

But both clinicians and patients are questioning the risks these new technologies may bring with them.

但臨床醫生和患者都在質疑這些新技術可能帶來的風險。

I think a lot of people are wondering, what's this going to look like when I go to the hospital, when I go to the doctor?

我想很多人都想知道，當我去醫院或看醫生時，會是什麼樣子？

How's it going to be used to inform decisions?

What about errors, false positive, false negatives?

誤差、假陽性、假陰性又是怎麼回事？

There's a lot of questions around this that really begs the expertise of people in different areas.

圍繞著這個問題，有很多問題需要不同領域的專家來解答。

Food and Drug Administration that does oversee certain uses of AI in healthcare.

美國食品和藥物管理局對人工ChatGPT在醫療保健領域的某些應用進行監督。

There are also groups like the American Medical Association and the guidelines and codes of conduct for responsible use of AI.

還有像美國醫學協會這樣的組織，以及負責任地使用人工ChatGPT的指導方針和行為準則。

So right now there are experts and there are various groups that are trying to oversee this and make sure that it's used responsibly.

現在有專家和各種團體正在努力監督這一點，並確保它得到負責任的使用。

So there's this big tension about how AI will affect our workforce.

關於人工智慧將如何影響我們的勞動力，存在著巨大的矛盾。

Clinicians have the ability to interact with patients, understand their preferences.

臨床醫生有能力與病人互動，瞭解他們的喜好。

So I still feel they'll have a key role to play in interpreting the outputs that the AI technology gives them.

我認為他們在解讀人工智慧技術提供的輸出結果方面仍將發揮關鍵作用。

While the larger AI labs from Google to OpenAI are trying to disrupt the healthcare industry at large, there are smaller and more focused labs within academia dedicated to applying AI to specialities.

從谷歌到 OpenAI 等規模較大的人工智慧實驗室正試圖顛覆整個醫療行業，而學術界也有一些規模較小、重點更突出的實驗室，致力於將人工智慧應用於專業領域。

Well, we assess how well the model performed looking at different use.

那麼，我們從不同的用途來評估模型的性能如何。

We're seeing an increasing investment in AI in healthcare.

我們看到，人工智慧在醫療保健領域的投資越來越多。

Since 2020, there has been a substantial increase in venture capital funding coming into this space.

自 2020 年以來，進入這一領域的風險投資資金大幅增加。

In 2024, we saw around $11 billion in funding going to various different AI and healthcare startups.

2024 年，我們看到約有 110 億美元的資金流向各種不同的人工智慧和醫療保健初創企業。

Experts say that academics have different priorities than some of these AI startups.

專家表示，學術界與其中一些人工智慧新創公司的優先事項不同。

Academics can fine tune large language models and they may have a different incentive to develop ethics guidelines around their responsible use.

學術界可以對大型語言模型進行微調，他們可能會有不同的動機來制定負責任地使用這些模型的道德準則。

I think our biggest risk with AI is making sure we use it responsibly and consistently.

我認為，人工智慧的最大風險在於確保我們以負責任的方式持續使用它。

We want people to pursue this creativity to the ends of the earth, but we also need clear processes in place to ensure that before they ever make it in front of a patient, they have been through a series of stage gates and reviews to ensure they are safe, effective, fair and equitable.

我們希望人們不遺餘力地追求這種創造力，但我們也需要制定明確的程序，以確保在它們出現在病人面前之前，經過一系列的階段性把關和審查，以確保它們是安全、有效、公平和公正的。

Yale has invested $150 million towards AI development over the next five years.

耶魯大學已投資 1.5 億美元，用於未來五年的人工智慧開發。

One of the ongoing efforts at Yale Medicine is the Cardiovascular Data Science Lab, founded by Dr. Rohan Khera in 2020.

心血管數據科學實驗室是耶魯大學醫學部正在開展的工作之一，該實驗室由羅漢-赫拉（Rohan Khera）博士於 2020 年創建。

So this is the Cardiovascular Data Science Lab, or the CARS Lab.

這就是心血管數據科學實驗室，簡稱 CARS 實驗室。

It's not a single domain of AI that we work on.

我們研究的不是單一的人工智慧領域。

We work on diagnostics, on precision care, or defining therapies that work for individuals, work for cardiac imaging and other ways that we use to enable better care for patients.

我們致力於診斷、精準醫療，或定義對個人有效的療法、心臟成像以及我們用來為患者提供更好護理的其他方法。

A good example is electrocardiograms, which are electrical recordings of the heart taken with stickers on the chest, available world over.

心電圖就是一個很好的例子，心電圖是用貼在胸口的貼紙對心臟進行的電記錄，全世界都有。

ECG-GPT, which is a play on ChatGPT, it takes an ECG image and generates a report.

ECG-GPT 致敬 ChatGPT，它獲取心電圖影像並生成報告。

這就是生成式人工智慧的一個例子。

We can generate a full report directly from the electrocardiogram.

我們可以直接從心電圖生成完整的報告。

Originally, this would require a cardiologist to interpret these studies, but our hope is that this can form the basis of very accurate reads that are available world over, that eventually clinicians can confirm and use in their care.

最初，這需要心臟科醫生來解讀這些研究，但我們希望這能為全世界提供非常準確的讀數奠定基礎，最終臨床醫生可以確認並在他們的護理中使用這些讀數。

So we decided to focus on app-based solutions, rather than things that integrate in large health systems, so that these tools are truly accessible.

我們決定專注於基於應用程式的解決方案，而不是整合到大型醫療系統中的東西，這樣這些工具才能真正為人們所用。

這是一張典型的 12 導聯心電圖。

And we know, me as a human reader, I can see these are the leads coming in from the chest.

我們知道，作為一個人類讀者，我可以看到這些是從胸腔裡出來的線索。

But for me looking at this, I could never tell you this is a patient with heart failure just based on their ECG.

但對我來說，我無法僅憑心電圖就判斷這是一名心衰患者。

What the AI eyeball is able to do is to kind of isolate these two leads and detect some signal that triggers that, yep, this is someone with left ventricular systolic dysfunction, an ejection fraction of less than 40%.

人工智慧眼球所能做的，就是隔離這兩個導聯，並檢測到一些信號，觸發「是的，這是一個左心室收縮功能障礙患者，射血分數低於 40%。」

And it visualizes it in a way that a human reader wouldn't be able to detect.

而且，它以人類讀者無法察覺的方式將其可視化。

While facilities like the Cards Lab are developing these new AI tools at a rapid pace, there are still many steps needed before these technologies can be implemented.

雖然像卡牌實驗室這樣的機構正在快速開發這些新的人工智慧工具，但要實現這些技術還需要很多步驟。

To ensure the public well-being, regulators may wish to impose oversight and guardrails before AI becomes the new standard practice of care.

為了確保公眾的福祉，監管機構不妨在人工智慧成為新的護理標準做法之前施加監督和警戒。

In terms of AI and misdiagnosis, experts say that the cases and examples of this are pretty limited at this point.

就人工智慧和誤診而言，專家表示，目前這方面的案例和實例還非常有限。

Because even though an AI might make a mistake or might make a misdiagnosis, there is still a real doctor involved in the diagnosis or prescription of a medication.

因為即使人工智慧可能會出錯或誤診，但在診斷或開藥方時，仍然有一位真正的醫生參與其中。

So are we really going to become healthier?

那麼，我們真的會變得更健康嗎？

I think it's not just the million-dollar question.

我認為這不僅僅是一個百萬美元的問題。

But it really requires that all of us, we need to be able to access, we need to be able to use this, we need to be willing to use it, and we need to have that space or energy to do it.

但是，這確實需要我們所有人，我們需要能夠獲取，我們需要能夠使用它，我們需要願意使用它，我們需要有這樣的空間或能量來做到這一點。

On the wall of my office, I have four photographs.

在我辦公室的牆上，掛著四張照片。

It's the transistor, the camera, the telephone, and the source code for the World Wide Web.

它是半導體、照相機、電話和網際網路的源代碼。

Because to me, those are the enabling innovations in my lifetime that transform the way we deliver care.

因為對我來說，在我有生之年，這些創新將改變我們提供醫療服務的方式。

Telemedicine, electronic health records, MRI imaging, all of that derives from those foundational innovations.

遠程醫療、電子病歷、核磁共振成像，所有這些都源於這些基礎創新。

我認為人工智慧就屬於這一類。