How AI and Democracy Can Fix Each Other | Divya Siddarth | TED

譯者: Lilian Chiu 審譯者: Shelley Tsang 曾雯海

Recently I told someone my work is on democracy and technology.

最近，我告訴一個人， 我的工作和民主及技術有關。

He turned to me and said,

他轉向我，說：「哇……

“Wow...

I’m sorry.”

我很遺憾。」

(Laughter)

（笑聲）

But I love my work.

但我喜歡我的工作。

I know that we can build a world

我知道我們可以建立這樣的 世界：在那裡，技術奇蹟

where technological marvels are directed towards people's benefit,

是以對人的益處為導向，

using their input.

並考量他們的意見。

We have gotten so used to seeing democracy as a problem to be solved.

我們已經非常習慣把民主 視為需要解決的問題。

但我把民主視為一種解決方案，

But I see democracy as a solution,

not as a problem.

而不是一個問題。

Democracy was once a radical political project,

民主曾經是個激進的政治計畫，

itself a cutting-edge social technology,

它本身就是一種尖端的社會技術，

是種新方法，

a new way to answer the very question we are faced with now,

可以解答我們現在在人工智慧 時代所面臨的這個問題：

in the era of artificial intelligence:

how do we use our capabilities to live well together?

我們要如何利用我們的能力

來好好共同生活？

We are told that transformative technologies like AI are too complicated,

別人告訴我們，轉型技術如人工智慧

太複雜了，或太危險了，

or too risky

or too important to be governed democratically.

或太重要了，因此不能 用民主方式來管理。

But this is precisely why they must be.

但這正是必須要用民主方式的原因。

If existing democracy is unequal to the task,

如果現有的民主無法勝任這項任務，

our job is not to give up on it.

我們的工作不是放棄它，

Our job is to evolve it.

我們的工作是去發展它，

And to use technology as an asset to help us do so.

並把技術當作資產來使用， 協助我們發展它。

Still, I understand his doubts.

然而，我能了解他的懷疑。

I never meant to build my life around new forms of democracy.

我從來沒有想過要把我的生活 建立在民主的新形式上。

I started out just really believing in the power of science.

我最初只是真的很相信科學的力量。

I was modifying DNA in my kitchen at 12,

十二歲時，我在廚房裡改造 DNA，

and when I got to Stanford as a computational biology major,

當我到史丹佛大學主修計算生物學，

I was converted to a new belief -- technology.

我改變了信仰，改信技術。

I truly believed in the power of tech to change the world.

我真心相信技術的力量 可以改變世界。

Maybe, like many of you.

也許在座許多人也如此相信。

But I saw that the technologies that really made a difference

但我看到，真正能造成改變的技術，

were the ones that were built with and for the collective.

是集體一起為了集體打造的技術。

Not the billions of dollars

不是投入到第十九個讓人上癮的 社群應用程式的數十億美元，

pumped into the 19th addiction-fueling social app.

But the projects that combine creating something truly new

而是像這樣的專案計畫： 真正創造新東西，同時

with building in ways for people to access,

在打造它時要考量讓大家能夠 取得它、從中受惠，並管理它。

benefit from and direct it.

Instead of social media, think of the internet.

想想網際網路，而不是社群媒體。

Built with public resources on open standards.

根據公開的標準， 用公共資源來打造。

This is what brought me to democracy.

這就是我進入民主領域的原因。

Technology expands what we are capable of.

技術會擴大我們的能力。

Democracy is how we decide what to do with that capability.

民主則是我們要如何決定 怎麼使用那些能力。

Since then, I've worked on using democracy as a solution

從那時起，我一直致力於 用民主來當解決方案，

in India, the US, the UK, Taiwan.

地點在印度、美國、英國、台灣。

I've worked alongside incredible collaborators

我和很棒的夥伴合作，

to use democracy to help solve COVID,

用民主來協助解決新冠肺炎疫情，

to help solve data rights.

協助解決資料權利。

And as I'll tell you today,

還有我今天要跟大家談的，

to help solve AI governance with policymakers around the world

協助解決人工智慧治理，

合作對象包括世界各地的政策 制訂者及走在尖端的科技公司，

and cutting-edge technology companies like OpenAI and Anthropic.

如 OpenAI 及 Anthropic。

How?

怎麼做？

By recognizing that democracy is still in its infancy.

要知道，民主還處在很初期的階段。

It is an early form of collective intelligence,

它是集體智慧的早期形式，

a way to put together decentralized input from diverse sources

是將來自各種來源的去中心化 輸入整合起來的方式，

and produce decisions that are better than the sum of their parts.

可以產生出比各部分 結合起來更好的決策。

That’s why, when my fantastic cofounder Saffron Huang and I

這就是為什麼當我 和我出色的共同創辦人黃珊

left our jobs at Google DeepMind and Microsoft

離開 Google DeepMind 和微軟的工作，

to build new democratic governance models for transformative tech,

為轉型科技打造 新的民主治理模型時，

I named our nonprofit the Collective Intelligence Project,

我把我們的非營利組織 命名為「集體智慧計畫」，

as a nod to the ever-evolving project of building collective intelligence

作為致敬，

向不斷發展且為了集體繁榮

而打造集體智慧的專案計畫致敬。

for collective flourishing.

Since then we've done just that,

從那時起，那就是我們在做的事，

building new collective intelligence models to direct artificial intelligence,

打造新的集體智慧模型， 來管理人工智慧，

to run democratic processes.

來執行民主流程。

And we've incorporated the voices of thousands of people into AI governance.

我們已將數千人的聲音 融入到人工智慧的治理中。

Here are a few of the things we've learned.

以下是我們學到的一些事。

First, people are willing and able to have difficult,

首先，

大家願意和能夠

進行困難、複雜的談話， 談微妙的主題。

complex conversations on nuanced topics.

When we asked people about the risks of AI they were most concerned about,

當我們問大家他們最關心 哪些人工智慧風險時，

they didn't reach for easy answers.

他們沒有只給簡單的答案。

Out of more than 100 risks put forward,

在被提出來的一百多項風險中，

the top-cited one: overreliance on systems we don't understand.

最多人引用的是：

過度依賴我們不理解的系統。

We talked to people across the country,

我們與全國各地的人交談，

from a veteran in the Midwest to a young teacher in the South.

從中西部的退伍軍人 到南方的年輕老師。

People were excited about the possibilities of this technology,

大家都對此技術的可能性感到興奮，

but there were specific things they wanted to understand

但對於模型能做些什麼， 他們會想要了解一些特定的資訊，

about what models were capable of before seeing them deployed in the world.

且要在模型被實際部署出去之前。

A lot more reasonable than many of the policy conversations that we're in.

比我們參與過的許多 政策談話都更明智有理。

And importantly, we saw very little of the polarization

重要的是，我們很少看到 我們老是聽到的兩極化，

we're always hearing about.

On average, just a few divisive statements

平均來說，

分裂的陳述和有共識的陳述比例 大約是幾則比上好幾百則。

for hundreds of consensus statements.

Even on the contentious issues of the day,

即使是當時極具爭議性的議題， 如自由言論，或種族與性別，

like free speech or race and gender,

we saw far more agreement than disagreement.

我們看到的一致意見 都遠多於不一致意見。

Almost three quarters of people agree that AI should protect free speech.

有近四分之三的人認同 人工智慧應保護自由言論。

Ninety percent agree that AI should not be racist or sexist.

九成的人認同人工智慧 不該有種族或性別歧視。

Only around 50 percent think that AI should be funny though,

不過，只有約一半的人認為 人工智慧應該要風趣點，

so they are still contentious issues out there.

這還是大家在爭議的議題。

These last statistics

最後這些統計數字

are from our collective constitution project with Anthropic,

出自我們與 Anthropic 合作的集體憲法專案計畫，

where we retrained one of the world's most powerful language models

在計畫中，我們重新訓練了 世界上最強大的語言模型之一，

on principles written by 1,000 representative Americans.

根據一千位有代表性的 美國人所寫的原則來訓練。

Not AI developers or regulators or researchers at elite universities.

不是人工智慧開發者或監管機構，

也不是菁英大學的研究人員，

We built on a way of training AI

我們立基的人工智慧訓練方式

that relies on a written set of principles or a constitution,

是以一組成文的原則或章程為基礎，

we asked ordinary people to cowrite this constitution,

我們請一般人來共同撰寫這份章程，

we compared it to a model that researchers had come up with.

我們將它與研究人員 提出的模型進行比較。

When we started this project, I wasn't sure what to expect.

我們開始這個計畫時， 我還不確定該抱什麼期待。

Maybe the naysayers were right.

也許反對者是對的。

AI is complicated.

人工智慧很複雜。

Maybe people wouldn't understand what we were asking them.

也許大家不會理解 我們要他們做什麼。

Maybe we'd end up with something awful.

也許我們最後的結果會很可怕。

But the people’s model, trained on the cowritten constitution,

但，大家的模型，根據 共同撰寫的章程訓練的模型，

was just as capable and more fair

能力沒有比較差，公平性還高於

than the model the researchers had come up with.

研究人員提出的模型。

People with little to no experience in AI

人工智慧方面的經歷為零或很少的人

did better than researchers, who work on this full-time,

做得比全職在此領域 工作的研究人員更好，

in building a fairer chatbot.

建立出更有公平性的聊天機器人。

Maybe I shouldn't have been surprised.

也許我不應該感到驚訝。

As one of our participants from another process said,

就如我們另一個流程的 一位參與者所言：

"They may be experts in AI, but I have eight grandchildren.

「他們也許是人工智慧專家， 但我有八個孫子孫女。

I know how to pick good values."

我知道如何挑選好的價值觀。」

If technology expands what we are capable of

如果技術擴大了我們的能力，

and democracy is how we decide what to do with that capability,

而民主是我們如何決定 怎麼使用那些能力，

here is early evidence that democracy can do a good job deciding.

那麼這裡就是些初期的證據， 證明民主能把決策工作做好。

Of course, these processes aren't enough.

當然，光這些流程還不夠。

Collective intelligence requires a broader reimagining of technology

集體智慧需要對技術及民主 做更廣泛的重新想像。

and democracy.

That’s why we’re also working on co-ownership models

這就是為什麼我們也投入 針對人工智慧背後的

for the data that AI is built on -- which, after all, belongs to all of us --

基礎資料建立共同擁有權模型——

畢竟，這些資料屬於我們所有人——

and using AI itself to create new and better decision-making processes.

並用人工智慧本身來創造 新的、更好的決策流程。

Taking advantage of the things

善用語言模型可以做到 而人類無法做到的事,

that language models can do that humans can’t,

like processing huge amounts of text input.

比如處理大量文字輸入資料。

Our work in Taiwan has been an incredible test bed for all of this.

我們在台灣做的研究是個很棒的 試驗床，為這一切奠定基礎。

Along with Minister Audrey Tang and the Ministry of Digital Affairs,

我們與唐鳳部長及數位發展部合作，

we are working on processes to ask Taiwan's millions of citizens

開發流程 ，來詢問 台灣的數百萬公民

他們希望看到怎樣的人工智慧未來。

what they actually want to see as a future with AI.

And using that input not just to legislate,

取得的意見不只會用來 立法，也會用來建造。

but to build.

Because one thing that has already come out of these processes

因為這些流程中已經 可以看出一個現象：

is that people are truly excited about a public option for AI,

大家對於人工智慧的公共選項 真的感到很興奮，

one that is built on shared public data that is reliably safe,

也就是建立在共享 公共資料上的選項，

這些資料可靠安全，

that allows communities to access, benefit from

社區可以取得、受惠於這些資料，

and adjust it to their needs.

並根據它們的需求調整這些資料。

This is what the world of technology could look like.

這就是技術世界可能的樣貌。

Steered by the many for the many.

由多數人為多數人掌舵。

I often find that we accept unnecessary trade-offs

我常發現，就轉型技術來說，我們會

when it comes to transformative tech.

接受不必要的權衡。

We are told that we might need to sacrifice democracy

別人告訴我們，我們 可能得要犧牲民主，

for the sake of technological progress.

為了技術進展，必須這麼做。

We have no choice but to concentrate power to keep ourselves safe

我們別無選擇，

只能集中力量，保護自己 免受可能的風險。

from possible risks.

This is wrong.

這是不對的。

It is impossible to have any one of these things --

不可能只有下列其中任何一項——

progress, safety or democratic participation --

進步、安全、民主參與——

without the others.

而不要另外兩項。

If we resign ourselves to only two of the three,

如果我們接受只要三項中的兩項，

we will end up with either centralized control or chaos.

最終的結果若不是 集中控制就是混亂。

Either a few people get to decide or no one does.

只有少數人能做決策， 不然就是沒人能做決策。

These are both terrible outcomes,

這兩種後果都很可怕，而我們的 研究顯示還有另一種方法。

and our work shows that there is another way.

Each of our projects advanced progress,

我們的每個專案計畫 都促進了進步、安全，

safety and democratic participation

以及民主參與，

by building cutting-edge democratic AI models,

做法是建立尖端的 民主人工智慧模型、

by using public expertise as a way to understand diffuse risks

用公眾專業知識來了解四散的風險，

and by imagining co-ownership models for the digital commons.

以及想像數位公地的 共同擁有權模型。

We are so far from the best collective intelligence systems we could have.

我們離我們能擁有的最佳 集體智慧系統還有好大的距離。

If we started over on building a decision-making process for the world,

如果我們重新開始 為世界建立一個決策流程，

what would we choose?

我們會選擇什麼？

Maybe we'd be better at separating financial power from political power.

也許我們會將金融權力

與政治權力分開得更好。

Maybe we'd create thousands of new models of corporations

也許我們會創造出數千種 企業或官僚制度的新模型。

or bureaucracies.

Maybe we'd build in the voices of natural elements

也許我們會融入自然元素 或未來世代的聲音。

or future generations.

Here's a secret.

說個小秘密。

In some ways, we are always starting from scratch.

在某些層面上，我們總是從零開始。

New technologies usher in new paradigms

新技術帶來了新的範式，

that can come with new collective intelligence systems.

這些範式可能會伴隨著 新的集體智慧系統。

We can create new ways of living well together

我們可以創造出讓大家 一起好好生活的新方法，

if we use these brief openings for change.

只要我們能善用 這些短暫的改變機會。

The story of technology and democracy is far from over.

技術和民主的故事離結束還很遠。

It doesn't have to be this way.

不見得要是這樣的。

Things could be unimaginably better.

一切可以更好許多。

As the Indian author Arundhati Roy once said,

如印度作家阿蘭達蒂‧ 羅伊曾經說過的：

"Another world is not only possible,

「另一個世界不僅是可能的，

she is on her way.

她正在路上了。

On a quiet day, I can hear her breathing."

在安靜的日子，我可以 聽到她的呼吸聲。」

I can hear our new world breathing.

我能聽到我們的新世界的呼吸聲。

One in which we shift the systems we have

在這個新世界裡，我們 會把我們擁有的系統轉化

towards using the solution of democracy to build the worlds we want to see.

成為民主的解決方案，

來建立我們想看到的世界。

The future is up to us.

未來取決於我們。

We have a world to win.

我們有一個世界要去贏取。

Thank you.

謝謝。

(Applause)

（掌聲）