A fascinating time capsule of human feelings toward AI Lucy FareyJones

I’m here, because I’ve spent
far too many nights lying awake,

worrying and wondering
who wins in the end.

Is it humans or is it robots?

You see, as a technology strategist,

my job involves behavior change:

understanding why and how
people adopt new technologies.

And that means I’m really frustrated

that I know I won’t live to see
how this all ends up.

And in fact, if the youngest person
watching this is 14

and the oldest, a robust 99,

then together,

our collective consciousnesses
span just 185 years.

That is a myopic pinprick of time

when you think of the evolution
and the story of life on this planet.

Turns out we’re all in the cheap seats

and none of us will live to see
how it all pans out.

So at my company,
we wanted a way around this.

We wanted to see if there was
a way to cantilever out,

beyond our fixed temporal vantage point,

to get a sense of how it all shakes up.

And to do this, we conducted a study
amongst 1,200 Americans

representative of the US census,

in which we asked a battery
of attitudinal questions

around robotics and AI

and also captured behavioral ones
around technology adoption.

We had a big study

so that we could analyze
differences in gender and generations,

between religious and political beliefs,

even job function and personality trait.

It is a fascinating,
time-bound time capsule

of our human frailty

in this predawn of the robotic era.

And I have five minutes
to tell you about it.

The first thing you should know
is that we brainstormed

a list of scenarios
of current and potential AI robotics.

They ran the spectrum from the mundane,

so, a robot house cleaner, anyone?

Through to the mischievous,

the idea of a robot pet sitter,
or maybe a robot lawyer,

or maybe a sex partner.

Through to the downright macabre,
the idea of being a cyborg,

blending human and robot,

or uploading your brain
so it could live on after your death.

And we plotted people’s comfort levels
with these various scenarios.

There were actually 31 in the study,

but for ease, I’m going to show you
just a few of them here.

The first thing you’ll notice,
of course, is the sea of red.

America is very uncomfortable
with this stuff.

That’s why we call it
the discomfort index,

not the comfort index.

There were only two things
the majority of America is OK with.

And that’s the idea
of a robot AI house cleaner

and a robot AI package deliverer,

so Dyson and Amazon, you guys should talk.

There’s an opportunity there.

It seems we’re ready to off-load
our chores to our robot friends.

We’re kind of definitely on the fence
when it comes to services,

so robot AI lawyer
or a financial adviser, maybe.

But we’re firmly closed
to the idea of robot care,

whether it be a nurse,
a doctor, child care.

So from this, you’d go,

“It’s OK, Lucy, you know what?

Go back to sleep, stop worrying,
the humans win in the end.”

But actually not so fast.

If you look at my data very closely,

you can see we’re more
vulnerable than we think.

AI has a branding problem.

So of those folks who said

that they would absolutely reject
the idea of a personal assistant,

45 percent of them had, in fact,
one in their pockets,

in terms of a device
with Alexa, Google or Siri.

One in five of those who were against
the idea of AI matchmaking

had of course, you guessed it,
done online dating.

And 80 percent of those of us
who refuse the idea

of boarding an autonomous plane
with a pilot backup

had in fact, just like me
to get here to Vancouver,

flown commercial.

Lest you think everybody
was scared, though,

here are the marvelous folk in the middle.

These are the neutrals.

These are people for whom you say,

“OK, robot friend,”

and they’re like,
“Hm, robot friend. Maybe.”

Or, “AI pet,”

and they go, “Never say never.”

And as any decent
political operative knows,

flipping the ambivalent middle
can change the game.

Another reason I know
we’re vulnerable is men –

I’m sorry, but men,
you are twice as likely than women

to believe that getting
into an autonomous car is a good idea,

that uploading your brain
for posterity is fun,

and two and a half times more likely
to believe that becoming a cyborg is cool,

and for this, I blame Hollywood.

(Laughter)

And this is where I want you
to look around the theater

and know that one in four men
are OK with the idea of sex with a robot.

That goes up to 44 percent
of millennial men

compared to just one in 10 women,

which I think puts a whole new twist
on the complaint of mechanical sex.

(Laughter)

Even more astounding
than that though, to be honest,

is this behavioral difference.

So here we have people who have
a device with a voice assistant in it,

so a smart speaker,
a home hub or a smart phone,

versus those who don’t.

And you can see from this graph

that the Trojan horse
is already in our living room.

And as these devices proliferate

and our collective defenses soften,

we all see how it can end.

In fact, this may be as good
a time as any to admit

I did take my Alexa Dot
on vacation with me.

Final finding I have time for
is generational.

So look at the difference
just three generations make.

This is the leap from silent
to boomer to millennial.

And what’s more fascinating than this
is if you extrapolate this out,

the same rate of change,

just the same pace,

not the accelerated one
I actually believe will be the case,

the same pace,

then it is eight generations away

when we hear every single American

thinking the majority
of these things here are normal.

So the year 2222 is an astounding place

where everything here is mainstream.

And lest you needed any more convincing,

here is the generation’s
“excitement level with AI.”

So not surprisingly,
the youngest of us are more excited.

But, and possibly the most
paradoxical finding of my career,

when I asked these people my 3am question,

“Who wins in the end?”

Guess what.

The more excited you are
about AI and robotics,

the more likely you are to say
it’s the robots.

And I don’t think we need a neural net
running pattern-recognition software

to see where this is all headed.

We are the proverbial frogs
in boiling water.

So if the robots at TED2222
are watching this for posterity,

could you send a cyborg, dig me up
and tell me if I was right?

(Laughter)

Thank you.

(Applause)

我在这里,因为我已经度过
了太多的夜晚,无法入睡,

担心并想知道
最终谁会赢。

是人类还是机器人?

你看,作为一名技术战略家,

我的工作涉及行为改变:

了解人们为什么以及如何
采用新技术。

这意味着我真的很沮丧

,因为我知道我不会活着
看到这一切如何结束。

事实上,如果最年轻的
人是 14 岁

,最年长的人是 99 岁

,那么

我们的集体意识
加起来只有 185 年。

当你想到
这个星球上的进化和生命故事时,这是一个短视的时间点。

事实证明,我们都坐在便宜的座位

上,我们谁都不会活着
看到这一切如何发展。

所以在我的公司,
我们想要解决这个问题。

我们想看看是否
有办法伸出悬臂,

超越我们固定的时间优势

,了解这一切是如何动摇的。

为此,我们对代表美国人口普查的 1,200 名美国人进行了一项研究,

我们在其中询问了一系列

有关机器人和人工智能的态度问题,

并捕捉
了有关技术采用的行为问题。

我们进行了一项大型研究,

以便我们可以
分析性别和世代、

宗教和政治信仰之间的差异,

甚至是工作职能和人格特征。 在机器人时代的黎明前,

这是一个迷人的、有
时限的时间胶囊

,体现了我们人类的脆弱

性。

我有五分钟的时间
告诉你。

您应该知道的第一件事
是,我们

集思广益,列出
了当前和潜在 AI 机器人的场景列表。

他们从平凡的范围内运行频谱,

所以,机器人房屋清洁工,有人吗?

通过恶作剧,

机器人宠物保姆的想法,
或者可能是机器人律师,

或者可能是性伴侣。

通过彻头彻尾的可怕,
成为一个机器人的想法,

将人类和机器人融合在一起,

或者上传你的大脑,
这样它就可以在你死后继续存在。

我们用这些不同的场景绘制了人们的舒适度

研究中实际上有 31 个,

但为了方便起见,我将在这里只向您展示
其中的一些。

当然,您首先会注意到
的是红色的海洋。

美国
对这些东西非常不舒服。

这就是为什么我们
称之为不适指数,

而不是舒适指数。

大多数美国人都可以接受的只有两件事。

这就是
机器人 AI 房屋清洁工

和机器人 AI 包裹递送员的想法,

所以戴森和亚马逊,你们应该谈谈。

那里有机会。

看来我们已经准备好把
我们的家务分担给我们的机器人朋友了。


服务方面,我们肯定是持谨慎态度的,

所以机器人 AI 律师
或财务顾问,也许吧。

但我们坚决
反对机器人护理的想法,

无论是护士
、医生还是儿童护理。

所以从这里,你会说,

“没关系,露西,你知道吗?

回去睡觉,别担心
,人类最终会赢。”

但实际上并没有那么快。

如果您仔细查看我的数据,

您会发现我们
比我们想象的更容易受到攻击。

人工智能存在品牌问题。

因此,在那些

表示他们绝对会拒绝
个人助理的想法的

人中,事实上,45% 的人口
袋里

有一个,就
带有 Alexa、谷歌或 Siri 的设备而言。 你猜

对了,在反对人工智能婚介的人中,有五分之一的人

当然
做过网上约会。

我们当中 80%

拒绝乘坐有飞行员后备的自动驾驶飞机的

人,事实上,就像我一样
去温哥华,

乘坐商业飞机。 不过,

为了避免你认为每个人
都害怕,

这里是中间的了不起的人。

这些是中立者。

你对这些人说,

“好吧,机器人朋友”

,他们就像,
“嗯,机器人朋友。也许吧。”

或者,“人工智能宠物”

,他们会说,“永不言败。”

正如任何体面的
政治人物所知道的那样,

翻转矛盾的中间人
可以改变游戏规则。

我知道
我们易受伤害的另一个原因是男性——

我很抱歉,但是男性,
你比

女性相信
进入自动驾驶汽车是个好主意,

为后代上传大脑很有趣的可能性是女性的两倍,

而且
相信成为机器人很酷的可能性要高出两倍半

,为此,我责怪好莱坞。

(笑声

) 这就是我想让
你环顾剧院的地方,

并知道四分之一的男人
对与机器人发生性关系的想法是可以的。 与十分之一的女性相比,

这一比例高达 44%
的千禧一代男性

我认为这对机械性行为的抱怨带来了全新的变化

(笑声)

说实话,比这更令人震惊的

是这种行为差异。

所以在这里,我们有一些人拥有
一个带有语音助手的设备,比如

智能扬声器
、家庭集线器或智能手机,

而那些没有的人。

从这张图可以

看出,
特洛伊木马已经在我们的客厅里了。

随着这些设备的激增

和我们集体防御的软化,

我们都看到了它会如何结束。

事实上,这可能
是承认

我确实带着我的 Alexa Dot
和我一起度假的好时机。

我有时间的最终发现
是世代相传的。

所以看看
仅仅三代人所产生的差异。

这是从沉默
到婴儿潮一代到千禧一代的飞跃。

比这更令人着迷的
是,如果你推断出来

,同样的变化率,

同样的速度,

而不是
我真正相信的加速

,同样的速度,

那么

当我们听到每一个 单身美国人

认为
这里的大多数事情都是正常的。

所以 2222 年是一个惊人的地方

,这里的一切都是主流。

为了避免你需要更多的说服力,

这里是这一代人的
“对人工智能的兴奋程度”。

所以毫不奇怪,
我们中最年轻的人更兴奋。

但是,这可能
是我职业生涯中最自相矛盾的发现,

当我凌晨 3 点问这些人我的问题时,

“最后谁赢了?”

你猜怎么着。


对人工智能和机器人技术越兴奋

,你就越有可能
说是机器人。

而且我认为我们不需要运行模式识别软件的神经网络

来了解这一切的发展方向。

我们是众所周知的沸水中的青蛙

所以,如果 TED2222 的机器人
正在为后代关注这个,

你能派一个机器人,
把我挖出来告诉我我是不是正确的?

(笑声)

谢谢。

(掌声)