artificial intelligence.

今天來到這裡，是想邀請大家 從一個新的角度來看我熟悉的領域：

I think the purpose of AI

人工智慧。

is to empower humans with machine intelligence.

我認為人工智慧的目的，

And as machines get smarter,

是要用機器智慧來增強人類的能力。

we get smarter.

而隨著機器變得更聰明，

I call this "humanistic AI" --

我們也會更聰明。

artificial intelligence designed to meet human needs

我稱這個為「人性化人工智慧」──

by collaborating and augmenting people.

透過與人工智慧合作並 強化人類的設計方式

Now, today I'm happy to see

來滿足人類的需求。

that the idea of an intelligent assistant

如今，我很高興能看到

is mainstream.

智慧助理的這個想法

It's the well-accepted metaphor for the interface between humans and AI.

能成為主流。

And the one I helped create is called Siri.

這象徵了人類與人工智慧之間的 互動介面已被廣泛接受。

You know Siri.

而其中一個 我協助創造的就是 Siri。

Siri is the thing that knows your intent

各位都認識 Siri。

and helps you do it for you,

Siri 會知道你的意圖，

helps you get things done.

並協助你

But what you might not know is that we designed Siri

把一切搞定。

as humanistic AI,

但各位可能不知道， 當初我們在設計 Siri 時，

to augment people with a conversational interface

就把它定位在 「人性化的人工智慧」，

that made it possible for them to use mobile computing,

透過對話介面的方式來強化人類，

regardless of who they were and their abilities.

讓使用行動計算成為可能，

Now for most of us,

不論他們是誰、 能力如何，都可以使用。

the impact of this technology

對我們大多數人而言，

is to make things a little bit easier to use.

受到這項技術的影響

But for my friend Daniel,

是讓我們在使用東西上 變得比較容易。

the impact of the AI in these systems is a life changer.

但對我朋友丹尼爾而言，

You see, Daniel is a really social guy,

卻是改變了他一生的人工智慧系統。

and he's blind and quadriplegic,

丹尼爾是個很愛社交的人，

which makes it hard to use those devices that we all take for granted.

他也是個四肢麻痺的盲人，

The last time I was at his house, his brother said,

我們認為理所當然的裝置設備， 對他來說卻很難用。

"Hang on a second, Daniel's not ready.

上次我去他家，他哥哥說：

He's on the phone with a woman he met online."

「等等，丹尼爾還沒準備好。

I'm like, "That's cool, how'd he do it?"

他正在和他線上認識的女生講電話。」

Well, Daniel uses Siri to manage his own social life --

我說：「好酷，他怎麼辦到的？」

his email, text and phone --

丹尼爾用 Siri 來管理 他自己的社交生活──

without depending on his caregivers.

他的電子郵件、訊息、電話──

This is kind of interesting, right?

不用依賴他的照護者。

The irony here is great.

這挺有趣的，對吧？

Here's the man whose relationship with AI

這也相當諷刺。

helps him have relationships with genuine human beings.

這個人與人工智慧的關係

And this is humanistic AI.

協助他與真正的人類建立了關係。

Another example with life-changing consequences

這就是人性化的人工智慧。

is diagnosing cancer.

還有一個可以 改變人生結果的例子，

When a doctor suspects cancer,

也就是癌症診斷。

they take a sample and send it to a pathologist,

當醫生懷疑你有癌症時，

who looks at it under a microscope.

他們會先取樣交給病理學家，

Now, pathologists look at hundreds of slides

病理學家再利用顯微鏡來看樣本。

and millions of cells every day.

病理學家每天要看數百片的載玻片

So to support this task,

及數百萬個細胞。

some researchers made an AI classifier.

所以為了支援這項工作，

Now, the classifier says, "Is this cancer or is this not cancer?"

有幾位研究人員做出了一種 人工智慧分類器。

looking at the pictures.

現在，分類器已經會看著圖片說：

The classifier was pretty good,

「這是不是癌症？」

but not as good as the person,

分類器挺厲害的，

who got it right most of the time.

但沒有人類那麼厲害，

But when they combine the ability of the machine and the human together,

人類大部份都能判讀正確。

accuracy went to 99.5 percent.

但，當機器與人類的能力 結合在一起時，

Adding that AI to a partnership eliminated 85 percent of the errors

正確率可以達到 99.5%。

that the human pathologist would have made working alone.

加上人工智慧的協作， 人類病理學家

That's a lot of cancer that would have otherwise gone untreated.

能減少個人判斷時 85% 的錯誤。

Now, for the curious, it turns out

這相當多啊，很多癌症患者， 當初就是因誤判而耽擱了治療。

that the human was better at rejecting false positives,

人類因為好奇的原因，所以

and the machine was better at recognizing those hard-to-spot cases.

比較擅長判斷錯誤的假陽性案例，

But the lesson here isn't about which agent is better

而機器比較擅長判斷 難以辨別的案例。

at this image-classification task.

但我們學到的並不是

Those things are changing every day.

在比誰在圖像分類上比較厲害。

The lesson here

那些東西每天都會變。

is that by combining the abilities of the human and machine,

這裡我們學到的是，

it created a partnership that had superhuman performance.

藉由結合人類與機器的能力，

And that is humanistic AI.

其創造出的合作關係 會有超人般的表現。

Now let's look at another example

而這就是人性化的人工智慧 。

with turbocharging performance.

我們再舉個例子，

This is design.

這方面的進步也是突飛猛進。

Now, let's say you're an engineer.

那就是「設計。」

You want to design a new frame for a drone.

假設你是個工程師，

You get out your favorite software tools, CAD tools,

你想要設計一個新的無人機骨架。

and you enter the form and the materials, and then you analyze performance.

你拿出你最愛的 軟體工具，電腦繪圖輔助工具，

That gives you one design.

你輸入了外形和材料， 然後按下分析表現。

If you give those same tools to an AI,

它只會跑出一種設計方案。

it can generate thousands of designs.

如果你把這些工具 拿給人工智慧使用，

This video by Autodesk is amazing.

它會幫你跑出數千種設計。

This is real stuff.

歐特克做的影片非常震撼。

So this transforms how we do design.

這是真實案例。

The human engineer now

這改變了我們設計的方式。

says what the design should achieve,

現在，人類工程師

and the machine says,

只要說出想要達成的設計，

"Here's the possibilities."

機器就會說：

Now in her job, the engineer's job

「有這些可能性方案。」

is to pick the one that best meets the goals of the design,

現在，她的工作，工程師的工作，

which she knows as a human better than anyone else,

就是選出一種最符合目標的設計，

using human judgment and expertise.

這是她身為人類最擅長的，

In this case, the winning form

用的就是人類的判斷和專業。

looks kind of like something nature would have designed,

在這案例中，脫穎而出的設計

minus a few million years of evolution

像是自然界設計出來的，

and all that unnecessary fur.

但這樣的設計， 不需要幾百萬年的演化

Now let's see where this idea of humanistic AI might lead us

也不需有那些皮毛。

if we follow it into the speculative beyond.

我們現在來想想， 如果我們跟著人性化人工智慧走，

What's a kind of augmentation that we would all like to have?

它會把我們帶到那些 意想不到的境界？

Well, how about cognitive enhancement?

我們想要的強化，有哪些方面？

Instead of asking,

認知強化如何？

"How smart can we make our machines?"

別再問：

let's ask

「我們能把機器做到多聰明？」

"How smart can our machines make us?"

我們要改問：

I mean, take memory for example.

「我們的機器能把我們變多聰明？」

Memory is the foundation of human intelligence.

我的意思是，例如「記憶力。」

But human memory is famously flawed.

記憶是人類智慧的基礎。

We're great at telling stories,

但人類的記憶力是出名的差。

but not getting the details right.

我們很會說故事，

And our memories -- they decay over time.

但無法把細節說正確。

I mean, like, where did the '60s go, and can I go there, too?

我們的記憶力會隨時間衰退。

(Laughter)

就像六十年代去哪了？ 我也能去那兒嗎？

But what if you could have a memory that was as good as computer memory,

（笑聲）

and was about your life?

但如果你能擁有和電腦 一樣好的記憶力，

What if you could remember every person you ever met,

而且這些記憶與你的人生有關， 那會如何呢？

how to pronounce their name,

如果你能記得你見過的每一個人、

their family details, their favorite sports,

叫得出每一個人的名字、

the last conversation you had with them?

家庭狀況的細節、他們最愛的運動、

If you had this memory all your life,

你和他們上次的對話，那會如何呢？

you could have the AI look at all the interactions

如果你一生都有這樣的記憶，

you had with people over time

你就可以要求人工智慧 去看看你過去

and help you reflect on the long arc of your relationships.

與人們的所有互動，

What if you could have the AI read everything you've ever read

並協助你反思你的人際關係弧線。

and listen to every song you've ever heard?

如果我們能讓人工智慧去閱讀 所有你閱讀過的東西、

From the tiniest clue, it could help you retrieve

去聽你聽過的每首歌，會如何呢？

anything you've ever seen or heard before.

從最微小的線索，它就能協助你找回

Imagine what that would do for the ability to make new connections

你以前看過或聽過的任何東西。

and form new ideas.

想像一樣，那對於

And what about our bodies?

建立新連結以及產生新點子的 能力會有什麼影響。

What if we could remember the consequences of every food we eat,

那麼，我們的身體呢？

every pill we take,

如果我們能記得 我們吃過的每樣食物、

every all-nighter we pull?

每顆藥物、

We could do our own science on our own data

記得每次熬夜 K 書的內容， 那會如何？

about what makes us feel good and stay healthy.

我們可以用自己的資料 來做自己的科學，

And imagine how this could revolutionize

找出什麼會讓我們 感覺很好並保持健康。

the way we manage allergies and chronic disease.

想像一下，這會為我們管理

I believe that AI will make personal memory enhancement a reality.

過敏及慢性病帶來怎樣的革命。

I can't say when or what form factors are involved,

我相信，人工智慧能夠 實現個人記憶的強化。

but I think it's inevitable,

我不知道何時或需要那些機緣，

because the very things that make AI successful today --

但我認為它是一定會到來的，

the availability of comprehensive data

因為人工智慧在現今 能夠成功的原因，

and the ability for machines to make sense of that data --

就是這些綜合性資料的可取得性，

can be applied to the data of our lives.

以及機器能從這些資料中 理解出意義的能力，

And those data are here today, available for all of us,

能為我們生活上的資料所用。

because we lead digitally mediated lives,

我們現今就已經能夠 取得那些資料了，

in mobile and online.

因為我們可以直接 用手機或上網的方式，

In my view, a personal memory is a private memory.

將我們的生活與數位科技做連結。

We get to choose what is and is not recalled and retained.

依我所見，個人的記憶 是私人的記憶。

It's absolutely essential that this be kept very secure.

我們可以選擇要或不要， 回憶或保留哪些記憶。

Now for most of us,

保持資料的安全絕對是有必要的。

the impact of augmented personal memory

對大部份人來說，

will be a more improved mental gain,

增強個人記憶希望產生的影響會是：

maybe, hopefully, a bit more social grace.

可以獲得心智上的成長

But for the millions who suffer from Alzheimer's and dementia,

或是希望可以有更高雅的社交能力。

the difference that augmented memory could make

但對於數百萬受阿茲海默症 及癡呆症所苦的人而言，

is a difference between a life of isolation

增加記憶能造成的改變，

and a life of dignity and connection.

就是讓那孤獨的人生，

We are in the middle of a renaissance in artificial intelligence right now.

變成有尊嚴及連結的人生。

I mean, in just the past few years,

我們現在正在人工智慧的 文藝復興時期當中，

we're beginning to see solutions to AI problems

僅僅在過去幾年，

that we have struggled with literally for decades:

我們就開始看到 人工智慧問題的解決方法，

speech understanding, text understanding,

這在過去數十年來， 一直是無法解決的問題，像是：

image understanding.

口語理解、文字理解、

We have a choice in how we use this powerful technology.

圖像理解。

We can choose to use AI to automate and compete with us,

我們可以選擇要 如何使用這強大的技術。

or we can use AI to augment and collaborate with us,

我們可以選擇用人工智慧 來做自動化並與我們競爭、

to overcome our cognitive limitations

或我們可以選擇用人工智慧 來做增強化並與我們合作，

and to help us do what we want to do,

來克服我們的認知限制，

only better.

並協助我們去做我們想要做的事，

And as we discover new ways to give machines intelligence,

而且做得更好。

we can distribute that intelligence

隨著我們發掘出新的方式 來將智慧賦予機器，

to all of the AI assistants in the world,

我們可以把那智慧移轉到

and therefore to every person, regardless of circumstance.

全世界所有的人工智慧助理身上，

And that is why,

且不論如何，都要確保 每個人都能因此受惠。

every time a machine gets smarter,

那就是為什麼

we get smarter.

每次當機器變更聰明，

That is an AI worth spreading.

我們也會變更聰明。

Thank you.

而這就是人工智慧值得推廣的原因。

(Applause)

謝謝大家。