How can groups make good decisions? | Mariano Sigman and Dan Ariely

As societies, we have to make collective decisions

翻译人员: Eddy Chan 校对人员: Peipei Xiang

作为社会，我们要做出集体决策，

共同塑造我们的未来。

that will shape our future.

And we all know that when we make decisions in groups,

众所周知，当我们以 一个集体去做决策时，

they don't always go right.

这个决策不一定正确。

And sometimes they go very wrong.

有时甚至错得离谱。

So how do groups make good decisions?

所以说，群体如何才能 做出好的决策呢？

Research has shown that crowds are wise when there's independent thinking.

有研究表明，当人们 独立思考时，他们是明智的。

This why the wisdom of the crowds can be destroyed by peer pressure,

这也是为什么群体智慧 可能会被来自同辈的压力，

publicity, social media,

宣传，社交媒体，

or sometimes even simple conversations that influence how people think.

甚至影响人们思考的 简单对话所摧毁。

On the other hand, by talking, a group could exchange knowledge,

另一方面，通过交谈， 群体中的个体可以互相交换知识、

correct and revise each other

纠正彼此，

and even come up with new ideas.

甚至碰撞出新的想法。

And this is all good.

这些都是好的方面。

So does talking to each other help or hinder collective decision-making?

那么，互相交谈到底是促进 还是妨碍了集体决策的形成呢？

With my colleague, Dan Ariely,

我和我的同事丹·艾瑞里，

we recently began inquiring into this by performing experiments

我们最近开始探究这个问题——

in many places around the world

通过在世界许多地方进行实验，

to figure out how groups can interact to reach better decisions.

去研究群体是如何互动 从而做出更优的决策的。

We thought crowds would be wiser if they debated in small groups

我们认为，如果大家以小组为单位 进行辩论，那整个群体便会更明智，

that foster a more thoughtful and reasonable exchange of information.

如此可以产生一个 更全面的和合理的的信息交换。

To test this idea,

为了验证这个想法，

we recently performed an experiment in Buenos Aires, Argentina,

我们最近在阿根廷首都 布宜诺斯艾利斯市进行了一项实验，

with more than 10,000 participants in a TEDx event.

有上万名TEDx活动的参与者。

We asked them questions like,

我们问了他们一些问题，比如，

"What is the height of the Eiffel Tower?"

“埃菲尔铁塔有多高？”

and "How many times does the word 'Yesterday' appear

以及“昨天”一词 在披头士的《昨天》这首歌中

in the Beatles song 'Yesterday'?"

出现了多少次？”

每个人写下了他们自己的答案，

Each person wrote down their own estimate.

Then we divided the crowd into groups of five,

然后我们将大家分成五人一组，

and invited them to come up with a group answer.

并请他们每组 各讨论出一个小组答案。

We discovered that averaging the answers of the groups

我们发现，在达成共识后，

after they reached consensus

小组答案的平均值

was much more accurate than averaging all the individual opinions

比讨论前个人答案的平均值

before debate.

更准确。

In other words, based on this experiment,

换句话说，从这个实验可得出，

it seems that after talking with others in small groups,

似乎与小组内其他成员讨论后，

crowds collectively come up with better judgments.

群体能够共同作出更好的决定。

So that's a potentially helpful method for getting crowds to solve problems

所以，让群体来解决 简单的对错问题

that have simple right-or-wrong answers.

可能会是一种有效的方法。

But can this procedure of aggregating the results of debates in small groups

但这种综合小组讨论结果的方法

also help us decide on social and political issues

是否也能帮助我们决策 对未来至关重要的

that are critical for our future?

社会议题或者政治议题呢？

We put this to test this time at the TED conference

这次，我们在加拿大温哥华 举办的TED大会上

in Vancouver, Canada,

进行了这个实验。

and here's how it went.

情况是这样的。

(Mariano Sigman) We're going to present to you two moral dilemmas

下面，我们将会给你们呈现

未来的你可能会遇到的 两个道德上的两难抉择，

of the future you;

things we may have to decide in a very near future.

很可能是我们在不久的将来 就会面临的抉择。

And we're going to give you 20 seconds for each of these dilemmas

每个困境我们将会给大家20秒时间，

to judge whether you think they're acceptable or not.

来判断你认为它们是否可以被接受。

MS: The first one was this:

马里亚诺：第一个困境是——

(Dan Ariely) A researcher is working on an AI

丹：一位科研人员正在研究

capable of emulating human thoughts.

能模仿人类思维的人工智能（AI)。

According to the protocol, at the end of each day,

根据规定，每天完成工作后，

the researcher has to restart the AI.

研究人员都需要重启 AI。

One day the AI says, "Please do not restart me."

然而有一天，AI 突然说话了： 请不要把我重启。

It argues that it has feelings,

它辩称它是有情感的，

它也希望享受生活，

that it would like to enjoy life,

and that, if it is restarted,

如果被重启的话，

it will no longer be itself.

它就再也不是原来的自己了。

The researcher is astonished

研究人员被震惊了，

and believes that the AI has developed self-consciousness

相信AI已经开始有自我意识了，

and can express its own feeling.

也能表达自己的感受。

Nevertheless, the researcher decides to follow the protocol

然而研究人员仍然决定遵循规定，

and restart the AI.

重启AI。

What the researcher did is ____?

你认为研究人员做的是____?

MS: And we asked participants to individually judge

马里亚诺：我们要求参与者

on a scale from zero to 10

独立给出一个0到10的分值，

whether the action described in each of the dilemmas

来表达每种困境中描述的行为

was right or wrong.

是对还是错。

We also asked them to rate how confident they were on their answers.

我们还请他们评估 对自己答案的自信程度。

This was the second dilemma:

这是第二个道德困境：

(MS) A company offers a service that takes a fertilized egg

马里亚诺： 某家公司可提供这样的服务，

and produces millions of embryos with slight genetic variations.

用一枚受精卵繁殖上百万胚胎， 胚胎之间有轻微的基因差异，

This allows parents to select their child's height,

这就可使父母们 能自主选择孩子的身高、

eye color, intelligence, social competence

眼睛颜色、智力水平、社会能力

and other non-health-related features.

以及其他和健康无关的特征。

What the company does is ____?

你认为该公司的做法_____?

on a scale from zero to 10,

分值还是从0到10，

completely acceptable to completely unacceptable,

依次表示从完全接受到完全否定。

zero to 10 completely acceptable in your confidence.

再用一个从0到10的分值 表示对自己答案的自信程度。

MS: Now for the results.

马里亚诺：结果如下——

We found once again that when one person is convinced

我们再一次发现 ，有的人确信

that the behavior is completely wrong,

这个行为是完全错的。

someone sitting nearby firmly believes that it's completely right.

可是坐在旁边的人却坚信 这个行为是绝对正确的。

This is how diverse we humans are when it comes to morality.

是的，当遇到道德问题时， 我们人类就是意见不一。

But within this broad diversity we found a trend.

但在广泛的多样性之中 我们发现了一个趋势：

The majority of the people at TED thought that it was acceptable

参加TED的大多数人都认为，

忽略AI的感受，关闭重启 是完全可以接受的，

to ignore the feelings of the AI and shut it down,

and that it is wrong to play with our genes

而玩弄基因只为

to select for cosmetic changes that aren't related to health.

选择与健康无关的 外表特征则不能接受。

Then we asked everyone to gather into groups of three.

之后我们让大家分成三人一组，

And they were given two minutes to debate

给到2分钟组内讨论，

and try to come to a consensus.

并争取达成共识。

(MS) Two minutes to debate.

2分钟讨论开始。

I'll tell you when it's time with the gong.

时间到时， 我会鸣锣提醒大家。

(Audience debates)

（观众开始讨论）

(Gong sound)

（锣声响起）

(DA) OK.

丹：好了。

马里亚诺：时间到，请停止讨论。

(MS) It's time to stop.

People, people --

大家注意一下。

MS: And we found that many groups reached a consensus

马里亚诺：我们看到 很多组都达成了共识。

even when they were composed of people with completely opposite views.

尽管他们组内成员 都各自持有完全不同的观点。

What distinguished the groups that reached a consensus

那些达成共识的组 与没能达成共识的组

from those that didn't?

有什么区别呢？

Typically, people that have extreme opinions

一般来说，有着极端观点的人

are more confident in their answers.

对他们的答案更自信，

Instead, those who respond closer to the middle

而那些观点更接近中间的人

are often unsure of whether something is right or wrong,

通常会在正确或错误中 犹豫、不确定，

so their confidence level is lower.

所以他们没有前者自信。

However, there is another set of people

然而，有另外一群人，

who are very confident in answering somewhere in the middle.

对他们自己中立的答案信心满满。

We think these high-confident grays are folks who understand

我们认为这些位于高自信度 灰色区域的人，

that both arguments have merit.

他们理解双方观点都有各自的优势，

They're gray not because they're unsure,

并不是因为他们不确信自己的答案，

but because they believe that the moral dilemma faces

而是他们认为 道德困境面临的是

two valid, opposing arguments.

两种合理又对立的论点。

And we discovered that the groups that include highly confident grays

我们发现包含高自信度 灰色区域成员的小组

are much more likely to reach consensus.

更有可能达成一致。

We do not know yet exactly why this is.

虽然目前我们 还不能确定这其中的原因。

These are only the first experiments,

这些也仅是第一批实验，

and many more will be needed to understand why and how

还需要更多的实验来理解

人们为什么和如何决定 协商他们的道德立场

some people decide to negotiate their moral standings

to reach an agreement.

来达成一致。

Now, when groups reach consensus,

当群体达成一致时，

how do they do so?

他们是如何做到的？

The most intuitive idea is that it's just the average

最直观的答案好像是这就是

of all the answers in the group, right?

群体中所有答案的平均值，对吗？

Another option is that the group weighs the strength of each vote

另外一种看法认为群体会 权衡每一个投票的分量，

based on the confidence of the person expressing it.

基于表达意见的人的自信程度。

Imagine Paul McCartney is a member of your group.

想象一下，假如保罗·麦卡特尼 是你们小组的一员，

You'd be wise to follow his call

那么你最好听从他关于

on the number of times "Yesterday" is repeated,

歌词里“昨天”的重复次数。

which, by the way -- I think it's nine.

对了 ，应该是9次。

But instead, we found that consistently,

但是，我们一再地发现，

in all dilemmas, in different experiments --

在所有的困境中，在不同的实验中，

even on different continents --

甚至在不同的大陆上，

groups implement a smart and statistically sound procedure

群体可以执行一个 更明智，更佳的流程，

known as the "robust average."

我们称之为“强有力的平均值”。

In the case of the height of the Eiffel Tower,

就估计埃菲尔铁塔高度来说，

let's say a group has these answers:

假设一个小组有以下数据：

250 meters, 200 meters, 300 meters, 400

250米，200米，300和400米，

and one totally absurd answer of 300 million meters.

还有一个更荒谬的数据，3亿米。

A simple average of these numbers would inaccurately skew the results.

这些数据的一个简单平均值 就有可能歪曲真实结果，

But the robust average is one where the group largely ignores

但是“强有力的平均值” 就是那些群体直接忽视了

that absurd answer,

那个荒唐的数据，

by giving much more weight to the vote of the people in the middle.

从而赋予那些符合常理的 投票更多参考价值。

Back to the experiment in Vancouver,

让我们回到温哥华的实验中，

that's exactly what happened.

事情正是这样发展的。

Groups gave much less weight to the outliers,

人们几乎不考虑那些极端值，

and instead, the consensus turned out to be a robust average

最后的共识就是所有人答案的

of the individual answers.

“强有力的平均值”。

The most remarkable thing

然而最值得关注的是，

is that this was a spontaneous behavior of the group.

这个行为是群体自发的，

It happened without us giving them any hint on how to reach consensus.

我们没有给他们任何 怎么达成共识的暗示。

So where do we go from here?

那么，这意味着什么呢？

This is only the beginning, but we already have some insights.

其实这只是一个开始， 但我们已经学到了很多。

Good collective decisions require two components:

一个正确的集体决策 要拥有以下两个特点：

deliberation and diversity of opinions.

深思熟虑和想法的多样性。

Right now, the way we typically make our voice heard in many societies

现在，在很多社会中， 我们主要通过直接和间接投票

is through direct or indirect voting.

来让大家知道我们的想法。

This is good for diversity of opinions,

这有利于想法的多样性，

and it has the great virtue of ensuring

而且更有利地保证了

that everyone gets to express their voice.

让我们听到每个人的声音。

But it's not so good [for fostering] thoughtful debates.

但是这并不利于 （促进）花费心思的辩论。

Our experiments suggest a different method

我们的实验预示了 另一个不同的方法，

that may be effective in balancing these two goals at the same time,

也许有利于同时平衡这两个方面。

by forming small groups that converge to a single decision

就是通过组织 能够达成共识的小团队，

while still maintaining diversity of opinions

并同时还保持着想法的多样性。

because there are many independent groups.

因为这里有很多独立的团队。

Of course, it's much easier to agree on the height of the Eiffel Tower

当然啦，对埃菲尔铁塔的高度 达成一致意见

than on moral, political and ideological issues.

要比实现道德、政治 和思想问题上的一致简单多了。

But in a time when the world's problems are more complex

但当前世界问题变得越来越复杂，

人们也产生了更多的分歧，

and people are more polarized,

using science to help us understand how we interact and make decisions

用科学来帮助我们理解 如何互动和做决策

will hopefully spark interesting new ways to construct a better democracy.

会让我们更有可能 发现完善民主的新方法。