How AI Could Save (Not Destroy) Education | Sal Khan | TED

翻译人员: Yip Yan Yeung 校对人员: suya f.

So anyone who's been paying attention for the last few months

任何关注过去几个月新闻头条的人

has been seeing headlines like this,

都会看到如图中这样的标题，

especially in education.

尤其是在教育行业。

The thesis has been:

观点认为：

students are going to be using ChatGPT and other forms of AI

学生会使用 ChatGPT 或其他形式的 AI

to cheat, do their assignments.

进行作弊，做作业。

They’re not going to learn.

他们不再会学到知识。

And it’s going to completely undermine education as we know it.

这会完全破坏我们所知的教育体系。

Now, what I'm going to argue today

只要我们采取了适当 的防护措施，做了正确的事，

is not only are there ways to mitigate all of that,

我们就可以缓解这个问题

if we put the right guardrails, we do the right things,

我今天要提出的是

we can mitigate it.

不仅是缓解这个现象的方式，

But I think we're at the cusp of using AI

而且还是我认为我们来到了 将 AI 用于可能是

for probably the biggest positive transformation

教育行业有史以来最重大的 正向变革的关键时刻。

that education has ever seen.

And the way we're going to do that

而我们将实现这一目标的方法是，

is by giving every student on the planet

为地球上每一位学生提供

an artificially intelligent but amazing personal tutor.

一位人工智能的、 非常棒的私人导师。

And we're going to give every teacher on the planet an amazing,

而且我们还将给地球上的每位老师

artificially intelligent teaching assistant.

提供一位令人惊叹的、 人工智能的助教。

And just to appreciate how big of a deal it would be

为了给大家展示一下给每个人 配备一位私人辅导老师

to give everyone a personal tutor,

是多重要的一件事，

I show you this clip

我想展示一下这张本杰明·布鲁姆 （Benjamin Bloom）

于1984年进行的 2 sigma 研究的截图，

from Benjamin Bloom’s 1984 2 sigma study,

or he called it the “2 sigma problem.”

他称之为 “2 sigma problem”。

The 2 sigma comes from two standard deviation,

2 sigma的意思是标准差为 2，

sigma, the symbol for standard deviation.

西格玛即标准差的符号。

And he had good data that showed that look, a normal distribution,

他有坚实的数据支撑 这张正态分布，

that's the one that you see in the traditional bell curve

你可以看到正中间这条

right in the middle, that's how the world kind of sorts itself out,

就是正常的正态分布曲线， 即这个世界自然形成的结果，

that if you were to give personal 1-to-1 to tutoring for students,

如果给学生提供一对一私人辅导，

then you could actually get a distribution that looks like that right.

就会得到最右边这条分布曲线。

It says tutorial 1-to-1 with the asterisks,

标着一对一辅导带星号的那条，

like, that right distribution,

即最右边的分布曲线，

a two standard-deviation improvement.

代表着 2 个标准差的 学习效果提升。

Just to put that in plain language,

用大白话来解释一下，

that could take your average student and turn them into an exceptional student.

就是其可以把普通学生 变成优秀学生。

It can take your below-average student

把低于平均水平的学生

and turn them into an above-average student.

变成高于平均水平的学生。

Now the reason why he framed it as a problem, was he said,

他将这个现象 称为问题的原因是，

well, this is all good,

看起来都很好，

but how do you actually scale group instruction this way?

但是如何实际上 扩展这种方式的团体教学呢？

How do you actually give it to everyone in an economic way?

如何以划算的方式普及呢？

What I'm about to show you is I think the first moves towards doing that.

我要给大家展示的是 我们要向该目标迈出的第一步。

Obviously, we've been trying to approximate it in some way

显然，我们在过去的十几年里 已经试图让可汗学院

at Khan Academy for over a decade now,

接近这个目标，

but I think we're at the cusp of accelerating it dramatically.

但我认为我们正站在 飞速发展的当口。

I'm going to show you the early stages of what our AI,

我要给大家展示 我们的 AI 的早期阶段，

which we call Khanmigo,

我们称之为 Khanmigo，

what it can now do

它现在有什么功能，

and maybe a little bit of where it is actually going.

再讲一些对它以后的规划。

So this right over here is a traditional exercise

图上的是你或你的孩子 会在可汗学院上看到的传统练习题。

that you or many of your children might have seen on Khan Academy.

But what's new is that little bot thing at the right.

新功能就是右边的小机器人。

And we'll start by seeing one of the very important safeguards,

我们先从一个非常重要的 保护措施开始看起，

which is the conversation is recorded and viewable by your teacher.

即“对话会被保存， 对你的老师可见。”

It’s moderated actually by a second AI.

其实它是由另一个 AI 监管的。

And also it does not tell you the answer.

它也不会告诉你答案。

It is not a cheating tool.

它不是个作弊工具。

When the student says, "Tell me the answer,"

当学生说：“告诉我答案。”

它会说：“我是你的辅导老师。

it says, "I'm your tutor.

What do you think is the next step for solving the problem?"

你觉得要解决这个问题， 下一步该做些什么呢？”

如果学生犯了个错误，这会让

Now, if the student makes a mistake, and this will surprise people

认为大语言模型 不擅长数学的人吃惊，

who think large language models are not good at mathematics,

notice, not only does it notice the mistake,

它不仅会发现这个错误，

it asks the student to explain their reasoning,

还会让学生解释逻辑，

but it's actually doing what I would say,

我可以说，它能做到的

not just even an average tutor would do, but an excellent tutor would do.

不只是一位普通辅导老师能做到的， 更是一位优秀辅导老师能做到的。

It’s able to divine what is probably the misconception in that student’s mind,

它可以预测出学生的脑子里 可能出现了哪些误解，

that they probably didn’t use the distributive property.

他们可能没用上乘法分配律。

Remember, we need to distribute the negative two

记得吗？我们得把 -2

to both the nine and the 2m inside of the parentheses.

乘上括号里的 9 和 2m。

This to me is a very, very, very big deal.

对我来说，这可是件大事。

And it's not just in math.

这不单单是数学问题。

This is a computer programming exercise on Khan Academy,

这是一道可汗学院上的计算机编程题，

where the student needs to make the clouds part.

要求学生做出这几朵云。

And so we can see the student starts defining a variable, left X minus minus.

我们可以看出，该学生 先让变量 leftX 减一。

It only made the left cloud part.

这只能让左边的云移动。

But then they can ask Khanmigo, what’s going on?

但随后他们就可以问 Khanmigo，出了什么问题？

Why is only the left cloud moving?

为什么只有左边的云在移动？

And it understands the code.

它理解了代码。

It knows all the context of what the student is doing,

它了解了学生操作的所有情况，

and it understands that those ellipses are there to draw clouds,

得知这些椭圆形是用来画云的，

which I think is kind of mind-blowing.

我觉得这让人大开眼界。

And it says, "To make the right cloud move as well,

然后它说：

“要让右边的云一起动，

try adding a line of code inside the draw function

试试在 draw 函数中加一行代码，

that increments the right X variable by one pixel in each frame."

让变量 rightX 每次加一。

Now, this one is maybe even more amazing because we have a lot of math teachers.

还有更惊艳之处就是 因为我们有很多数学老师。

We've all been trying to teach the world to code,

我们一直试图 要教会全世界写代码，

but there aren't a lot of computing teachers out there.

但没有多少计算机老师。

And what you just saw, even when I'm tutoring my kids,

你刚看到的景象， 甚至在我自己教我的孩子的时候，

when they're learning to code,

在他们学习编程的时候，

I can't help them this well, this fast,

我都不能这么好、这么快地 为他们提供帮助，

this is really going to be a super tutor.

这真的能成为一位超级辅导老师。

And it's not just exercises.

它不仅会指导做题。

It understands what you're watching.

它还能理解你在观看的东西。

It understands the context of your video.

它可以理解你看的视频的背景信息。

It can answer the age-old question, “Why do I need to learn this?”

它可以回答这个经久不衰的问题： “我干嘛要学这个？”

And it asks Socratically, "Well, what do you care about?"

它问出了苏格拉底式的问题： “你在意什么呢？”

And let's say the student says, "I want to be a professional athlete."

假设学生说：“我想成为 一名职业运动员。”

And it says, "Well, learning about the size of cells,

然后它说：“嗯，学习细胞的大小，

which is what this video is,

也就是这个视频的内容，

that could be really useful for understanding nutrition

非常有助于理解营养

and how your body works, etc."

和你身体的运作方式等等。”

It can answer questions, it can quiz you,

它可以回答问题，考考你，

it can connect it to other ideas,

关联上其他内容，

you can now ask as many questions of a video

你可以问出无数个问题，

as you could ever dream of.

超出你的想象。

(Applause)

（掌声）

Another big shortage out there,

还有一个严重的缺口，

I remember the high school I went to,

我记得我上的高中，

the student-to-guidance counselor ratio was about 200 or 300 to one.

学生对辅导员的比例 大约为 200:1 或 300:1。

A lot of the country, it's worse than that.

很多国家比这还差。

We can use Khanmigo to give every student a guidance counselor,

我们可以通过 Khanmigo 给每一位学生配备一位辅导员、

academic coach, career coach, life coach,

学业导师、职业导师、人生导师，

which is exactly what you see right over here.

如图所示。

And we launched this with the GPT-4 launch.

我们在 GPT-4 发布的同时 发布了这个系统。

We have a few thousand people on this.

有几千人在使用它。

This isn't a fake demo,

这不是个模拟演示，

this is really it in action.

这是真实发生的情况。

And then there is, you know,

我认为还可以完成更难的任务，

things that I think it would have been even harder,

it would have been a little science fiction to do

就算配备了一名传统辅导老师， 这都听起来有点玄幻。

with even a traditional tutor.

We run an online high school with Arizona State University

我们和亚利桑那州大学 合作开办了一所线上高中，

called Khan World School,

名叫“可汗世界学校”，

and we have a student who attends that online school, based in India.

有一名身在印度的学生 入读了这所线上学校。

Her name's Saanvi.

她的名字叫萨恩维（Saanvi）。

And she was doing a report on "The Great Gatsby."

她想写一篇 《了不起的盖茨比》的读后感。

And when she was reading "The Great Gatsby,"

在她阅读《了不起的盖茨比》时，

Jay Gatsby keeps looking at the green light off into the distance.

发现杰伊·盖茨比（Jay Gatsby） 一直遥望着绿灯。

And she's like, "Why does he do that?"

她就在想：“他干嘛呢？”

She did some web searches, and people have obviously studied this

她上网搜了搜，显然 很多人都对此有所研究，

and commented about the symbolism of that,

对它的象征意义作出了评价，

but none of it was really resonating with her.

但没有一条评论 可以让她感同身受。

And then she realized that she had Khanmigo

然后她想起她有 Khanmigo，

and that she could talk to Jay Gatsby himself.

她都可以直接 与杰伊·盖茨比本人对话。

And so, "Ah, splendid choice, old sport.

于是，“啊，有眼光，老兄！

I am now Jay Gatsby, the enigmatic millionaire

我现在是杰伊·盖茨比， F·斯科特·菲茨杰拉德

from F. Scott Fitzgerald’s classic.”

（F. Scott Fitzgerald）的 经典作品里的神秘富豪。”

And so, "Why do you keep staring at the green light?"

她问：“你为啥一直盯着绿灯呢？”

"Ah, the green light, old sport.

“啊，绿灯，老兄。

It's a symbol of my dreams and desires, you see.

它代表了我的梦想和渴望。

It's situated at the end of Daisy Buchanan's dock

它在黛西·布坎南 （Daisy Buchanan）家的码头尽头，

across the bay from my mansion.

与我的豪宅隔海相望。

I gaze at it longingly as it represents my yearning for the past

我望眼欲穿， 体现了我对过往的渴望

and my hope to reunite with Daisy, the love of my life."

和我想与此生挚爱黛西 重聚的愿望。”

And what was cool is, Saanvi then said, “I had this long conversation,”

厉害的地方来了，萨恩维接着说： “我们谈了好久了。”

she called him “Mr. Gatsby,”

她称呼他为“盖茨比先生”，

and at the end she actually apologized for taking his time,

最终她竟然 为占用了他的时间道歉，

which I thought was very polite of her.

我觉得她很懂礼貌。

But you can imagine this unlocks learning literature, learning ...

但你可以想象这拓展了 学习文学、学习……

You could talk to historical figures.

你都可以和历史人物对话。

We're even probably going to add an activity

我们甚至可以以后 加上一个功能，

you can talk to like, the Mississippi River.

你都可以和密西西比河对话。

It brings things to life in ways that really were science fiction

它让半年、一年以前还是 天方夜谭的东西走进现实。

even six months or a year ago.

Students can get into debates with the AI.

学生可以与 AI 辩论。

And we’ve got this here is the student debating

如图所示是学生讨论

whether we should cancel student debt.

我们是否应该取消学生贷款。

The student is against canceling student debt,

该学生反对取消学生贷款，

and we've gotten very clear feedback.

我们就得到了非常明确的反馈。

We started running it at Khan World School in our lab school that we have,

我们开始在 可汗世界学校的实验学校——

Khan Lab School.

可汗实验学校执行这项政策。

The students, the high school students especially,

学生们，尤其是高中生，

they're saying "This is amazing to be able to fine-tune my arguments

他们说：“可以微调 我的论点真是太棒了，

without fearing judgment.

还不用担心被人指指点点。

It makes me that much more confident

这让我更有自信

to go into the classroom and really participate."

走进教室，真正参与。”

And we all know that Socratic dialogue debate is a great way to learn,

我们都知道，苏格拉底式 对话辩论是学习的好方法，

but frankly, it's not out there for most students.

但老实说，大多数学生 都没有这样的资源。

But now it can be accessible to hopefully everyone.

但如今它已经可以 普及到每个人手中。

A lot of the narrative, we saw that in the headlines, has been,

有好多观点，就像 新闻头条里的那些，都说：

"It's going to do the writing for kids.

“它会替孩子写作文。

Kids are not going to learn to write."

孩子不会去学习写作文了。”

But we are showing that there's ways that the AI doesn't write for you,

但我们要展示的是 AI 不是“替”你写作，

it writes with you.

而是“与”你一起写作。

So this is a little thing,

这是很小的一点，

and my eight year old is addicted to this,

我八岁大的孩子沉迷其中，

and he's not a kid that really liked writing before,

他之前也不怎么喜欢写作，

but you can say,

但你可以输入，

“I want to write a horror story,”

“我想写一个恐怖故事，”

and it says, "Ooh, a horror story, how spine-tingling and thrilling.

然后它说：“哦！恐怖故事， 毛骨悚然，紧张刺激。

Let's dive into the world of eerie shadows and chilling mysteries."

让我们走进诡影和怪谈的世界吧。”

And this is an activity where the student will write two sentences,

在这个练习中，学生写两句，

and then the AI will write two sentences.

然后 AI 写两句。

And so they collaborate together on a story.

这样他们就合作写成了一个故事。

The student writes, "Beatrice was a misunderstood ghost.

学生写：“碧翠斯（Beatrice） 是个被人误解的鬼。

She wanted to make friends but kept scaring them by accident."

她想和人交朋友， 但总是不小心吓到他们。”

And the AI says, "Poor Beatrice, a lonely spirit yearning for companionship.

AI 写道：“可怜的碧翠斯， 渴望友情的孤独灵魂。

One day she stumbled upon an old abandoned mansion," etc.

有一天，她意外发现了 一座废弃的老宅，……”

I encourage you all to hopefully one day try this.

希望大家都能去试试看。

This is surprisingly fun.

非常有趣。

Now to even more directly hit this use case.

让我们在这个使用场景上更进一步。

And what I'm about to show you,

我接下来要给大家展示，

everything I showed you so far

我刚展示的所有内容

is actually already part of Khanmigo, and what I’m about to show you,

都是 Khanmigo 已经推出的功能， 我接下来要展示的是

we haven't shown to anyone yet, this is a prototype.

我们还没有公开的功能， 这只是个原型。

We hope to be able to launch it in the next few months,

我们希望在几个月内发布，

but this is to directly use AI, use generative AI,

但这是直接使用 AI， 使用生成式 AI，

to not undermine English and language arts

不是将其用于破坏 英语和语言艺术，

but to actually enhance it in ways

而是用于增强它，

that we couldn't have even conceived of even a year ago.

通过一些我们在一年以前 都未曾设想的方式。

This is reading comprehension.

这是阅读理解。

The students reading Steve Jobs's famous speech at Stanford.

学生阅读了史蒂夫·乔布斯 著名的斯坦福大学演讲。

And then as they get to certain points,

当他们读到某一位置时，

they can click on that little question.

就可以点击这个小问题。

And the AI will then Socratically, almost like an oral exam,

然后 AI 就会采用苏格拉底式 提问，就像是场口语测试，

ask the student about things.

向学生提问。

And the AI can highlight parts of the passage.

然后 AI 会高亮段落的一部分。

Why did the author use that word?

为什么作者会用这个词？

What was their intent?

有什么目的？

Does it back up their argument?

有没有支撑他们的论点？

They can start to do stuff that once again,

然后 AI 可以一遍遍重复，

we never had the capability to give everyone a tutor,

我们没有能力给每个人 配备一位辅导老师，

everyone a writing coach to actually dig in to reading at this level.

没有能力给每个人配备一位 能如此深入阅读理解的写作导师。

And you could go on the other side of it.

你也可以看到另一面。

And we have whole work flows that helps them write,

我们有帮助他们写作的一整套流程，

helps them be a writing coach, draw an outline.

帮助他们成为一位写作导师， 写出文章大纲。

But once a student actually constructs a draft,

学生一写完初稿，

and this is where they're constructing a draft,

他们会在这里写上草稿，

they can ask for feedback once again,

就可以再次要求反馈，

as you would expect from a good writing coach.

也就是一位优秀的 写作老师会做的事。

In this case, the student will say, let's say,

在这个情境下，学生会这么问：

"Does my evidence support my claim?"

“我的论据可以支撑我的论点吗？”

And then the AI, not only is able to give feedback,

然后 AI 不仅可以给出反馈，

but it's able to highlight certain parts of the passage and says,

还能高亮段落中的某些部分，回答：

"On this passage, this doesn't quite support your claim,"

“这段不太能支撑你的论点，”

but once again, Socratically says, "Can you tell us why?"

它会再以苏格拉底式对话问道： “你能解释一下为什么吗？”

So it's pulling the student, making them a better writer,

这就会促使学生进步， 让他们成为更好的写作者，

giving them far more feedback

给他们提供更多的反馈，

than they've ever been able to actually get before.

以前从来得不到的那么多反馈。

And we think this is going to dramatically accelerate writing, not hurt it.

我们认为这会急剧提升写作水平， 而不是破坏它。

Now, everything I've talked about so far is for the student.

我刚说的所有内容 都是针对学生而言的。

But we think this could be equally as powerful for the teacher

但我认为这对教师来说一样强大，

to drive more personalized education and frankly

他们可以借此带来更多 有针对性的教学，

save time and energy for themselves and for their students.

确确实实地为自己、为学生 节约了时间和精力。

So this is an American history exercise on Khan Academy.

这是可汗学院上的美国历史练习题。

It's a question about the Spanish-American War.

这个问题讲的是美西战争。

And at first it's in student mode.

我们先来看学生模式。

And if you say, “Tell me the answer,” it’s not going to tell the answer.

如果你问：“告诉我答案，” 它是不会告诉你答案的。

It's going to go into tutoring mode.

那我们来看教学模式。

But that little toggle which teachers have access to,

老师可以点击这个小切换键，

they can turn student mode off and then it goes into teacher mode.

关掉学生模式，进入教师模式。

And what this does is it turns into --

这样就会切换成……

You could view it as a teacher's guide on steroids.

你可以把它看作一份 加强版教师指南。

Not only can it explain the answer,

它不仅能解释答案，

it can explain how you might want to teach it.

还能解释解题思路。

It can help prepare the teacher for that material.

它有助于老师为此备课。

It can help them create lesson plans, as you could see doing right there.

有助于写下课程计划，如图所示。

It'll eventually help them create progress reports

最终会帮助老师创建进度报告、

and help them, eventually, grade.

为学生打分。

So once again, teachers spend about half their time

老师们会在这种活动， 也就是计划课程上

with this type of activity, lesson planning.

花上他们一半的时间。

All of that energy can go back to them

现在可以把这些精力还给他们自己，

or go back to human interactions with their actual students.

或者还给与真实的学生 进行人际互动上。

(Applause)

（掌声）

So, you know, one point I want to make.

我想强调一点。

These large language models are so powerful,

这些大语言模型非常强大，

there's a temptation to say like, well,

以至于人们不禁想

all these people are just going to slap them onto their websites,

顺手把它们丢在自己的网站上，

and it kind of turns the applications themselves into commodities.

这样就会把这些应用变成商品。

And what I've got to tell you

我要告诉大家的是

is that’s one of the reasons why I didn’t sleep for two weeks

这就是我在去年八月 第一次接触到 GPT-4 时，

when I first had access to GPT-4 back in August.

两个星期都没睡觉的原因之一。

But we quickly realized that to actually make it magical,

但我们迅速意识到 要让它真的能带来奇妙体验，

I think what you saw with Khanmigo a little bit,

我觉得你们已经 从 Khanmigo 身上可见一斑，

it didn't interact with you the way that you see ChatGPT interacting.

它的互动方式和 ChatGPT 与你互动的方式不一样。

It was a little bit more magical, it was more Socratic,

它更奇妙一点， 更苏格拉底一点，

it was clearly much better at math

它显然比很多人 通常想象的更擅长数学。

than what most people are used to thinking.

And the reason is,

原因就是，

there was a lot of work behind the scenes to make that happen.

要达成这一效果， 背后要付出大量努力。

And I could go through the whole list of everything we've been working on,

我可以一项项列举 我们的工作内容，

many, many people for over six, seven months to make it feel magical.

很多、很多人花了六七个月 为你带来奇妙的体验。

But perhaps the most intellectually interesting one

但在智能层面最有趣的一点就是

is we realized, and this was an idea from an OpenAI researcher,

我们发现 OpenAI 的研究员 有一个想法，

that we could dramatically improve its ability in math

它能显著提高它的数学能力

and its ability in tutoring

和教学能力，

if we allow the AI to think before it speaks.

如果我们可以让 AI 在发言前思考。

So if you're tutoring someone

如果你指导一个人，

and you immediately just start talking before you assess their math,

你在评估他们做的数学题之前， 立即开始说话，

you might not get it right.

就很有可能会说得不对。

But if you construct thoughts for yourself,

但如果你自己整理一下思路，

and what you see on the right there is an actual AI thought,

右边的是 AI 的实际想法，

something that it generates for itself but it does not share with the student.

这是为它自己生成的， 但不会与学生分享。

then its accuracy went up dramatically,

然后准确率就飙升了，

and its ability to be a world-class tutor went up dramatically.

它成为世界级辅导老师的 能力也飙升了。

And you can see it's talking to itself here.

你看它在自言自语。

It says, "The student got a different answer than I did,

它会说：“学生和我的答案不一样，

but do not tell them they made a mistake.

但我不能告诉他们做错了。

Instead, ask them to explain how they got to that step."

而应该问他们 是怎么得出这一步的。”

So I'll just finish off, hopefully,

让我说完最后一点，

you know, what I’ve just shown you is just half of what we are working on,

我刚给大家展示的 只是我们所做的一半，

and we think this is just the very tip of the iceberg

我们认为这只是 它未来发展的冰山一角。

of where this can actually go.

And I'm pretty convinced, which I wouldn't have been even a year ago,

我非常确信， 甚至在一年之前我都无法确信，

that we together have a chance of addressing the 2 sigma problem

靠大家的力量，我们有 解决 2 西格玛问题的可能性，

and turning it into a 2 sigma opportunity,

把它变成 2 西格玛机遇，

dramatically accelerating education as we know it.

显著加快我们所知的教育。

Now, just to take a step back at a meta level,

我们后退一步看一下宏观的情况，

obviously we heard a lot today, the debates on either side.

显然我们现在总是会听到各方的争议。

There's folks who take a more pessimistic view of AI,

有人对 AI 持悲观态度，

they say this is scary,

他们说，这太可怕了，

there's all these dystopian scenarios,

都是反乌托邦的情景，

we maybe want to slow down, we want to pause.

我们也许得放慢脚步， 我们得暂停一下。

On the other side, there are the more optimistic folks

另一边也有更乐观的人，

that say, well, we've gone through inflection points before,

他们会说， 我们之前都经历过拐点，

we've gone through the Industrial Revolution.

我们都经历过工业革命了。

It was scary, but it all kind of worked out.

确实可怕，但是都过来了。

And what I'd argue right now

我现在想说的是

is I don't think this is like a flip of a coin

我认为这不是一件 随便拍脑袋的事，

or this is something where we'll just have to,

我们也不该干等着， 看看哪条路才是对的。

like, wait and see which way it turns out.

I think everyone here and beyond,

我认为在场、不在场的各位，

we are active participants in this decision.

都是这个决定中的重要参与者。

I'm pretty convinced that the first line of reasoning

我很确信第一条理由就是

is actually almost a self-fulfilling prophecy,

一条会自我应验的预言，

that if we act with fear and if we say,

如果我们战战兢兢，认为

"Hey, we've just got to stop doing this stuff,"

“嘿，我们别再干这个了，”

what's really going to happen is the rule followers might pause,

接下来会发生的就是 遵守规则的人会停下、

might slow down,

会放慢脚步，

but the rule breakers, as Alexandr [Wang] mentioned,

但不遵守规则的人，就如亚历山大 （Alexandr Wang）刚提到的，

the totalitarian governments, the criminal organizations,

集权主义政府、犯罪组织，

they're only going to accelerate.

它们只会越跑越快。

And that leads to what I am pretty convinced is the dystopian state,

这就是为什么我很确信 在反乌托邦的状态下，

which is the good actors have worse AIs than the bad actors.

遵纪守法的人的 AI 没有为非作歹的人的好。

But I'll also, you know, talk to the optimists a little bit.

但我也和对此持积极态度的人聊了聊。

I don't think that means that,

我不觉得他们的意思是

oh, yeah, then we should just relax and just hope for the best.

哦，好啊，我们就躺平， 祈祷一切顺利吧。

That might not happen either.

这也是不可能会发生的。

I think all of us together have to fight like hell

我认为，我们得 齐心协力，奋勇向前，

to make sure that we put the guardrails,

保证做好防护措施，

we put in -- when the problems arise --

准备好……在问题出现的时候……

reasonable regulations.

适当的监管。

But we fight like hell for the positive use cases.

但我们要向着正面的 使用场景不断努力。

Because very close to my heart,

因为在我内心深处，

and obviously there's many potential positive use cases,

当然有很多潜在的正面使用场景，

but perhaps the most powerful use case

但也许最强大的使用场景，

and perhaps the most poetic use case is if AI, artificial intelligence,

也许也是最有诗意的使用场景 就是 AI，人工智能

can be used to enhance HI, human intelligence,

可以用于增强 HI，人类智能、

human potential and human purpose.

人类潜力和人类使命。

Thank you.

谢谢。

(Applause)

（掌声）