人工智能是否能用于医疗系统

Dr Siegel, so I went to the dentist the other day and he did his normal inspection then he did the AI inspection .

西格尔医生，前几天我去看牙医，他给我做了常规检查然后做了人工智能检查。

And there wasn't much different but there were some things that he missed that the AI was able to pick up.

没有太大的不同，但他遗漏了一些人工智能能够识别的东西。

So is this a good thing or bad thing?

所以这是好事还是坏事?

So that part is a good thing.

这是件好事。

That's kind of like an automatic pilot on a plane but do you want automatic pilot landing the plane, you don't, we've been on a plane together .

这有点像飞机上的自动驾驶仪，但你想让自动驾驶仪降落飞机吗，你不想，我们都坐过飞机。

We certainly don't want that, we want to pilot in their LJ and we also have to be scared that Jeffrey Hinton quit google today.

我们当然不想这样，甚至想穿救生衣，杰弗里·辛顿现在退出了谷歌，我们也不得不担心。

Because and he's the father, he's the godfather of deep learning he started all of this decades ago.

因为他是深层学习的教父，他在几十年前就开始了这一切。

And he quit, saying this has gotten out of control.

然后他辞职，说这已经失控了。

You already said at the beginning where it stays in control, that tool you talked about at Harvard is called Civil and it looks at a CAT scan, a lung CAT scan of a smoker and it says before a radiologist would ever notice this, this could be lung cancer, this spot worries me.

你一开始就说过它在掌控一切，你讲过哈佛的那个工具叫做西维尔，它看着分层造影扫描，一个吸烟者的肺部分层造影扫描，它在放射科医生注意到这个之前说，这可能是肺癌，这个斑点让我担心。

And also you can use AI to look at things that you didn't even think of looking at with it, with a radio geologist.

你也可以用人工智能来观察那些你甚至没有想过用它来观察的东西，和一个无线电地质学家一起。

To begin with, so it's a tool, where it gets out of control is where somebody says that doctor didn't have enough time for me.

首先，它是一个工具，它失控的时候是有人说医生没有时间给我看病。

Let's go to AI, let's go to ChatGPT and ChatGPT says “you're feeling fine, sir” and I'm thinking and then study out of San Diego says that shows more empathy than I do.

让我们去找人工智能，去问ChatGPT，ChatGPT说，先生，你感觉很好，然后圣地亚哥的研究表明这比我更有同理心。

I got to work on my empathy then.

那我得培养我的同理心了。

We have to go to doctors for the human element, for the personalized medicine, for the back and forth your case is different than another person's case.

我们必须去看医生，因为人性的因素，因为个性化的药物，因为你的情况和其他人的情况不同。

AI won't be able to figure that out.

人工智能是搞不清楚的。

It just seems to be a liability in the making, I mean it's a robot and eventually it's going to miss something.

它似乎只是一个在制造中的不利因素，我的意思是，它是一个机器人，它终会遗漏一些东西。

I understand using it as a second pair of eyes to say, okay you asked me to infer on this maybe I need a different X-ray.

我能理解把它作为第二只眼睛用，然后说，好吧，想要推断这个我需要不同的X光片。

But to just go to a visit of a robot I don't know if I could ever do that, Doc.

但是我不知道我能不能去找一个机器人看病，医生。

You just made a really great point if a medical center relies on AI and the AI gets it wrong, there's even a term for this, it's called hallucinating.

你刚才说得很好，如果医疗中心依赖人工智能，而人工智能做错了，这甚至有一个术语，叫做幻觉。

AI can hallucinate, can you believe that, like suddenly AI doesn't know where it is and it gets something wrong.

人工智能会产生幻觉，你相信吗，就像突然间，人工智能不知道它在哪里，它会出错。

The medical center is still liable and how about me versus AI making a diagnosis, what if AI gets something right and I get it wrong, I'm liable.

医疗中心还是有责任的，我和人工智能做的诊断相比，如果人工智能做对了，而我做错了，我就有责任。

All of this liability and the thing that bothers me the most is insurance companies using AI to determine if you get that test approved or not.

所有这些责任和最困扰我的事情是保险公司使用人工智能来确定你是否获得了测验的批准。

I can't argue that against the computer.

我不能反驳电脑。

Also the cost, okay if I'm paying for a doctor, I'm not going to pay AI the same thing for a computer to do it.

还有成本，如果我请医生看病，我不会给人工智能同样的钱让电脑来看病。

It just doesn't make sense, it's not jiving with me, Doc.

这根本说不通，它不适合我，医生。

I agree with that AI could have a union, you have to pay us more, robots will align, this is where it's heading.

我同意人工智能可以有一个工会，你必须给它们更多的工资，机器人将结盟，这是它的发展方向。

But seriously as long as we keep it as a tool that we doctors use, that nurses use, that medical centers use for research, for contributing to diagnoses, I'm fine with it, I don't want it making decisions.

但认真地说，只要我们把它作为一种工具，给医生使用，护士使用，医疗中心使用，进行研究，为诊断做出贡献，我就可以接受它，我不希望它做决定。

Also why did you guys go to all that schooling, if this was going to be the case.

还有，如果是这样的话，你们为什么要去上那么多学。

Well it's scary that AI actually did better on some of these tests than I could.

好吧，在这些测试中，人工智能实际上比我做得更好，这很可怕。

Is this the end of your profession?

你的职业生涯就此结束了吗?

No I'm saying that it's a wake-up call that we got to re-look at our empathy, how we make complex decisions and the art of medicine.

不，我是说，这是一个警钟，让我们重新审视我们的同理心，如何做出复杂的决定，以及医学的艺术。

We've got to keep to that.

我们必须坚持这一点。

That's the most important thing, AI cannot do the art of medicine.

这是最重要的一点，人工智能无法了解医学的艺术。

Doc, I think you find it alarming but I think a lot of people are buying into and that's scary.

医生，我想你觉得这很令人担忧，但很多人都在入股人工智能，这很可怕。

Thank you so much for joining the program, I hope you keep us posted.

非常感谢你加入我们的节目，希望你能和我们保持联系。