How a Worm Could Save Humanity From Bad AI | Ramin Hasani | TED

翻译人员: Yip Yan Yeung 校对人员: Bruce Wang

My wildest dream is to design artificial intelligence

我最大胆的梦想就是设计出 可以和我们做朋友的人工智能。

that is our friend, you know.

If you have an AI system that helps us understand mathematics,

如果有一个帮助我们 理解数学的 AI 系统，

you can solve the economy of the world.

你就能解决世界经济问题。

If you have an AI system that can understand humanitarian sciences,

如果有一个能够理解 人道主义科学的 AI 系统，

we can actually solve all of our conflicts.

我们就能解决所有冲突。

I want this system to, given Einstein’s and Maxwell’s equations,

我希望这个系统能够利用 爱因斯坦和麦克斯韦的方程组

take it and solve new physics, you know.

得出新的物理学。

If you understand physics, you can solve the energy problem.

理解了物理学， 你就能解决能量问题。

So you can actually design ways for humans

你可以为人类设计出

to be the better versions of themselves.

成为更好自己的方法。

I'm Ramin Hasani.

我是拉明·哈萨尼 （Ramin Hasani）。

I’m the cofounder and CEO of Liquid AI.

我是 Liquid AI 的联合创始人 兼首席执行官。

Liquid AI is an AI company built on top of a technology

Liquid AI 是一家基于我在 麻省理工学院发明的技术的 AI 公司。

that I invented back at MIT.

It’s called “liquid neural networks.”

这项技术被称为“液态神经网络”。

These are a form of flexible intelligence,

它是一种灵活的智能形式，

as opposed to today's AI systems that are fixed, basically.

对应如今多为固定的 AI 系统。

So think about your brain.

想想你的大脑。

You can change your thoughts.

你可以改变你的想法。

When somebody talks to you,

有人和你说话时，

you can completely change the way you respond.

你可以完全改变你的回应方式。

You always have a mechanism that we call feedback in your system.

你的系统中总是有一种 我们称之为“反馈”的机制。

So basically when you receive information from someone as an input,

当你从某人那里收到信息 作为输入时，

you basically process that information, and then you reply.

你会处理这些信息，然后回复。

For liquid neural networks,

在液态神经网络中，

we simply got those feedback mechanisms, and we added that to the system.

我们就有这种反馈机制， 然后将其添加到系统中。

So that means it has the ability of thinking.

意味着它具有思维能力。

That property is inspired by nature.

这项技术的灵感来自大自然。

We looked into brains of animals and, in particular,

我们研究了动物的大脑，特别是

a very, very tiny worm called “C. elegans”

一种名为“秀丽隐杆线虫”的 极微小蠕虫。

The fascinating fact about the brain of the worm

蠕虫的大脑有一个有趣的事实，

is that it shares 75 percent of the genome that it has with humans.

它有 75% 与人类大脑的基因组相同。

We have the entire genome mapped.

我们已经绘制了整个基因组。

So we understand a whole lot

我们对它的神经系统的 功能也了解很多。

about the functionality of its nervous system as well.

So if you understand the properties of cells in the worm,

如果你了解蠕虫细胞的特性，

maybe we can build intelligent systems that are as good as the worm

也许我们可以打造 和蠕虫匹敌的智能系统，

and then evolve them into systems that are better than even humans.

再把它们改良成 甚至优于人类的系统。

The reason why we are studying nature is the fact that we can actually,

我们之所以研究自然， 是因为我们可以

having a shortcut through exploring all the possible kind of algorithms

避免探索力所能及设计出的 所有算法，找到捷径。

that you can design.

You can look into nature,

你可以研究大自然，

that would give you like, a shortcut

它会给你一条捷径，

to really faster get into efficient algorithms

更快地找到高效的算法，

because nature has done a lot of search,

因为大自然已经进行了大量的搜索，

billions of years of evolution, right?

经历了数十亿年的进化，对吧？

So we learned so much from those principles.

我们从这些原理中学到了很多东西。

I just brought a tiny principle from the worm

我刚刚将蠕虫带来的一个小原理

into artificial neural networks.

引入了人工神经网络。

And now they are flexible,

这样它们就很灵活，

and they can solve problems in an explainable way

可以用前所未有 可解释的方式解决问题。

that was not possible before.

AI is becoming very capable, right?

AI 变得非常强大了，对吧？

The reason why AI is hard to regulate

AI 之所以难以监管，

is because we cannot understand, even the people who design the systems,

是因为我们无法理解， 即使是设计系统的人，

we don't understand those systems.

我们都不了解这些系统。

They are black boxes.

它们是个黑盒。

With Liquid, because we are fundamentally using mathematics

有了 Liquid，由于我们其实 是在利用可以理解的数学，

that are understandable,

we have tools to really understand

我们有了工具，真正理解

and pinpoint which part of the system is responsible for what,

和确定系统的哪一部分有什么作用，

you're designing white box systems.

那就是在设计白盒系统。

So if you have systems that you can understand their behavior,

如果你有可以理解其行为的系统，

that means even if you scale them into something very, very intelligent,

就意味着即使你将它们拓展为 非常、非常智能的系统，

you can always have a lot of control over that system

你始终可以因为理解该系统 而对它有高度的掌控。

because you understand it.

You can never let it go rogue.

你不会让它失控。

So all of the crises we are dealing with right now,

我们现在面临的所有危机，

you know, doomsday kind of scenarios,

比如世界末日之类的情景，

is all about scaling a technology that we don't understand.

都是关于拓展一项我们不了解的技术。

With Liquid, our purpose is to really calm people down

借助 Liquid，我们的目标 是真正让人们冷静下来，

and show people that, hey,

向人们表明，嘿，

you can have very powerful systems,

你可以拥有非常强大的系统，

that you have a lot of control and visibility

你可以高度掌控、清晰看见 它们的运行机制。

into their working mechanisms.

The gift of having something [with] superintelligence is massive,

拥有超级智能有大量的好处，

and it can enable a lot of things for us.

它能让许多事成真。

But at the same time,

但同时，

we need to have control over that technology.

我们必须掌控这项技术。

Because this is the first time that we’re going to have a technology

因为这是我们第一次 拥有一项技术，

that is going to be better than all of humanity combined.

它会比我们全人类更加优秀。