The Real-World Danger of Online Myths | Vidhya Ramalingam | TED

譯者: Lilian Chiu 審譯者: SF Huang

"You are a disgusting liar."

「你是個讓人作嘔的騙子。」

"Someone, somewhere will hunt you down."

「在某處一定會有某人 絕對不會放過你的。」

"I hope someone puts a bullet between your eyes."

「我希望有人一槍斃了你。」

These are messages received by climate scientists.

這些是氣候科學家收到的訊息。

According to a recent survey,

根據最近的一項調查，

39 percent of climate scientists have faced online abuse.

39% 的氣候科學家 都曾遇過網路暴力。

18 percent of those are threats of physical violence.

其中 18％ 是人身暴力相關的威脅。

"At the end of the day,

「到頭來，

we're going to see just how much you believe

我們就來看看你有 多相信你的全球暖化，

in your global warming

and whether you're willing to die for your so-called 'research.'"

以及你是否願意為了 你所謂的『研究』而死。」

No scientist should have to fear for their lives.

科學家不應該還得要 擔心自己的生命安全。

But this is just another day in the life of a climate scientist.

但這卻只是氣候科學家 稀鬆平常的一天。

I'm not a climate scientist.

我不是氣候科學家。

I'm not a climate change activist.

我不是氣候變遷活動家。

I'm a counterterrorism expert.

我是反恐專家。

I started my journey meeting with white supremacists

我這段旅程的開端，

是在瑞典的地下室 和白人至上主義者會面。

in basements in Sweden

and went on to lead a global policy effort

接著，我領導了一項全球政策行動，

after Europe's first major terrorist attack

時間是在白人至上主義者第一次

perpetrated by a white supremacist.

向歐洲發動重大恐怖攻擊之後。

I went on to found Moonshot,

我接著成立了「宏大目標」，

an organization that works to end violence online.

這個組織的宗旨 是致力終結網路暴力。

I care about climate change denial

我很在乎對於氣候變遷的否認，

because it's so often weaponized

因為它經常被當成武器來使用，

to serve as a justification for violence.

為暴力行為找正當理由。

It would be easy to think

很容易就會認為，

that if only we could get people to understand climate change is real,

如果我們能讓大家了解 氣候變遷是真實的，

we could put an end to this.

我們就能消除這種狀況。

Unfortunately, it's not that simple.

不幸的是，沒那麼簡單。

In 2019,

2019 年，

a gunman walked into a Walmart in El Paso, Texas.

一名槍手走進德州 厄爾帕索的沃爾瑪。

He killed 23 people,

他殺死了二十三個人，

many of immigrant background.

許多都有移民背景。

He called himself an “ecofascist.”

他自稱「生態法西斯主義者」。

He believed in climate change,

他相信氣候變遷，

but he had bought into mis- and disinformation

但他誤信了錯誤資訊和假資訊，

that immigrants were the root cause of it,

以為氣候變遷是移民造成的，

that sustainability would only be possible

且唯有消滅有色人種， 才有可能實現永續。

with the elimination of people of color.

Mis- and disinformation are so often weaponized

錯誤資訊和假資訊 太常被當成武器來用，

成為使用暴力的正當理由。

to serve as a justification for violence.

Although they're often used interchangeably,

儘管這兩個詞常被 當成是一樣的意思，

misinformation is information that's false or misleading.

但錯誤資訊指的是不正確 或會誤導人的資訊。

Disinformation is spread intentionally to cause harm.

假資訊則是為了造成傷害 而蓄意傳播出去的資訊。

It’s so powerful because it taps into your grievances,

它非常強大，因為 它能觸及你的不滿，

what makes you really angry,

讓你非常憤怒，

and it offers simplistic solutions.

並提供過度簡化的解決方案。

There's typically a villain and a hero.

通常都有一個反派和一個英雄。

Over the last two years,

在過去兩年間，我和我的團隊 一直在研究世界各地中，

my team and I have been researching different kinds of manipulation tactics

used all over the world to spread disinformation.

用來散播假資訊的各種操弄策略。

Two of the most common were decontextualization

其中最常見的兩種策略是脫離情境

and fearmongering.

和製造恐慌。

Decontextualization is the practice of taking information

脫離情境的做法是讓資訊 脫離它原本的情境，

out of its original context to deliberately mislead people.

刻意去誤導別人。

For example, earlier this year,

例如，今年早些時候，

Europe experienced a series of protests by farmers

歐洲發生了一連串的農民抗爭，

against a range of proposed environmental regulations.

反對提議出來的一系列環境規定。

There were street blockades and protests,

有街道封鎖、抗議、 示威、佔領等活動。

demonstrations, occupations.

Adding to an already tense moment,

當時局勢已經夠緊張了， 在火上再加油的是

several inauthentic images circulated.

有數張不真實的圖像四處流傳。

This one purported to show

這張圖片聲稱位在 巴黎的烏克蘭大使館

the Ukrainian embassy in Paris getting pummeled with manure.

遭到糞便攻擊。

This was actually footage taken months earlier

這其實是在幾個月前拍攝的，

from an entirely different protest

是一場完全不同的抗議活動，

about an entirely different issue in Dijon,

反對的議題也不同，

not even in Paris.

地點甚至不是巴黎，而是第戎。

And this effort to mislead the public,

而這種誤導公眾的行動，

it wouldn't be complete without the use of new technology.

若沒有使用新技術就不可能辦到。

Here's an image showing the streets of Paris lined with bales of hay.

這張圖片顯示巴黎的 街道堆滿了乾草捆。

It's a really striking image, isn't it?

這張圖很引人注目，對吧？

This never happened.

從來沒發生過這種事， 這張圖完全是由人工智慧生成的。

It was entirely generated by AI.

And this isn't just happening in Europe.

且這種情況不只出現在歐洲。

Last year,

去年，

after wildfires raged in Hawaii,

在夏威夷遭野火肆虐之後，

a disinformation network linked to the Chinese Communist Party

一個與中國共產黨有關的假資訊網絡

spread inauthentic images

散播了不真實的圖像，

purporting that the US government had intentionally spread the wildfires

聲稱美國政府用所謂的「氣象武器」

using a so-called “weather weapon.”

蓄意引發這些野火。

Can you imagine?

各位能想像嗎？

Over a hundred people died in those wildfires,

有超過一百人因為那些野火喪命，

竟有人認為野火是他們 自己的政府蓄意引發的，

and the idea that those fires were deliberately set

by their own government

against their own people?

用來對抗自己的人民？

It's terrifying.

太駭人了。

這類陰謀論的說法 可能引發廣泛的恐懼，

These kinds of conspiratorial narratives can spread widespread fear,

which takes us to the next powerful tactic of disinformation:

此時就要談到第二種 強大的假資訊策略：

fearmongering:

製造恐慌：

deliberately exaggerating an issue

蓄意把議題誇大，

so that you can provoke fear and alarm.

目的在引發恐懼和警覺心。

We know that emotion-driven information processing

我們知道，由情緒 驅動的資訊處理方式

can overtake evidence-based decision making,

可以取代以證據為基礎的決策，

which is what makes this form of disinformation so effective.

這就是這種形式的 假資訊如此有效的原因。

It's for these reasons that a recent MIT study found

基於這些理由，麻省理工學院 近期的一項研究發現

a false story will travel six times quicker

假故事傳播給一千五百個人的速度

to reach 1,500 people than a true story will.

是真實故事的六倍。

And we know Facebook fact-checkers take up to 72 hours

我們知道臉書的事實查核員 平均要花七十二小時

on average

to identify and remove this content.

才能辨識出這些內容並將之刪除。

By that time, most impressions have already been made.

等到刪除時，錯誤的印象 早就已經建立起來了。

Now I know we have all seen this online,

我知道大家都在網路上看過這些，

and when you see it happen,

看到假資訊時，你就會 很想要用事實來回應。

it can be really tempting to respond with the facts.

I get it.

我懂。我們以邏輯和科學為榮。

We pride ourselves on logic and science.

The truth matters.

真相很重要。

So when someone is so obviously spreading false information,

所以當很明顯有人在散播假資訊時，

just correct them, right?

提出糾正就行了，對吧？

Unfortunately, this doesn't always work.

不幸的是，這麼做不見得行得通。

Believe me, I spent the last two decades

相信我，過去二十年來我一直在學習

learning how to have conversations with people buying into white supremacy.

如何與相信白人比較優越的人交談。

That is disinformation at its worst.

那是最惡劣的假資訊。

Disinformation wins

假資訊能得逞，

because of the emotions it inspires,

是因為它能激發出的情緒，

because of the way it makes people feel.

是因為它能讓人產生的感受。

So if someone is so bought into disinformation,

因此，如果某人對假資訊深信不疑，

getting into debates on the facts with them

和他們爭辯事實反而可能 會把他們進一步推到角落，

can just risk pushing them even further into a corner,

backing them into a corner

他們被逼到角落後，

so that they get really defensive.

就會豎起極高的防備心。

OK, so if we can't debate the facts endlessly,

好，如果我們不能無止盡地 不斷爭辯事實，那能怎辬辦？

what can we do?

Last year, Moonshot partnered with Google

去年，「宏大目標」與 Google 合作

to test an approach known as “prebunking.”

測試一種稱為「預先闢謠」的方法。

Prebunking is a proven communication technique

預先闢謠是一種 已通過驗證的溝通技巧，

designed to help people spot and reject efforts

旨在幫助大家發現和抵制

to manipulate them in the future

未來試圖操弄他們的作為，

by giving them forewarning

做法是預先警告他們，

and giving them tools to be able to reject a manipulative message,

並提供他們可以抵制 操弄訊息的工具。

you lessen the likelihood that they will be misled.

這樣就能減少他們被誤導的機會。

This is not about telling people what is true or false

這並不是要告訴大家什麼 是真的或假的，是對的或錯的。

or right or wrong.

It's about empowering people to protect themselves.

重點在於讓大家 有能力可以保護自己。

We've tapped into the universal human desire not to be manipulated,

我們利用的是人類普遍 不希望被操弄的慾望，

and this method has been tried and tested for decades, since the 1960s.

且這種方法從 1960 年代起 就經過數十年的反覆測試。

All prebunking messages contain three essential ingredients.

所有要做預先闢謠的訊息 都包含三個基本要素。

One:

第一：

an emotional warning.

情感上的警告。

You alert people that there are others out there

你要警告大家，外頭有些人

who may be trying to mislead or manipulate them.

可能會試圖誤導或操弄他們。

Be aware, you may be targeted.

請注意，你可能會成為目標。

Two:

第二：

stimulus.

刺激。

You show people examples of manipulative messaging

你給大家看一些操弄式訊息的例子，

so that they will be more likely

這樣他們未來就比較 有可能辨識出這類訊息。

to be able to identify those in the future.

And three: refutation.

第三：反駁。

You give people the tools

你提供工具給大家， 讓他們能即使反駁訊息。

to be able to refute a message in real time.

For example,

比如，

if you see a headline that’s really sensational,

如果你看到一個能引起轟動的標題，

and it either seems too good to be true or it makes you really angry,

也許是太好了，好到不像真的， 也許是會讓你非常憤怒，

always Google around for other sources.

務必一定要上網搜尋其他來源。

Always Google around.

一定要上網搜尋。

OK, so we knew the steps to take,

好，所以，我們知道該採取的步驟，

but we also knew if we were going to really get at this problem

但我們也知道，如果我們真的 要解決世界各地的這個問題，

around the world,

不可能靠單一種通用的方法就辦到。

a one-size-fits-all approach wouldn't work.

We knew we needed to get local.

我們知道在地化是必要的。

So we partnered with civil society organizations

因此，我們與世界各國的 公民社會組織合作，

in countries around the world,

from Germany to Indonesia to Ukraine.

從德國、到印尼、到烏克蘭。

And we started first with the evidence.

我們先從證據著手。

We met with dozens of experts,

我們與數十位專家會面，

我們對線上環境做了調查，

we surveyed the online space,

and we identified the most common manipulation tactics

並找出每個國家最常 被使用的操弄戰略。

being used in each country.

We then partnered with local filmmakers

接著，我們與當地電影製作人合作，

to create educational videos

製作教育影片，來教導大家

that would teach people about those manipulation tactics

和他們所屬國家常用的 操弄手法的相關資訊，

that were being used in their home country.

In some contexts,

在某些情境中，

we found that people trust close peers and relatives the most.

我們發現大家最相信的 是親密的同儕和親人。

So in Germany,

所以，在德國，

we filmed close friends chatting in a park.

我們拍攝的是親密的 朋友在公園裡聊天。

In Ukraine,

在烏克蘭，

we filmed family dialogues around a kitchen table,

我們拍攝的則是全家 坐在餐桌前的對話，

a setting that's so familiar to so many of us,

那是許多人都很熟悉的場景，

where so many of us have had those difficult conversations.

許多人都經歷過必須要進行 這類困難談話的情況。

We wanted to encourage people to have these kinds of conversations

我們想鼓勵大家在自己的 信任圈內進行這類談話，

within their own trusted circles,

whether they're in El Salvador or Indonesia.

不論是在薩爾瓦多或印尼。

And to do so before pivotal moments

且要在關鍵時刻之前就做，

where online manipulation efforts intensify,

關鍵時刻指的是線上操弄 越演越烈的時刻，

like elections.

比如選舉時。

So as we prepared to head into the EU elections,

在歐盟選舉即將到來之前，

we knew that distrust in climate science

我們知道對氣候科學的不信任

had already emerged as a critical misinformation theme.

已經成為了重要的錯誤資訊主題。

Now one study had found that adults over the age of 45

一項研究發現， 四十五歲以上的成年人

are less likely to investigate false information

在網路上碰到假資訊時， 比較不會去查證它。

when they stumble across it online.

Now we also know that adults over the age of 45

我們也知道四十五歲以上的成年人

have higher voter turnout,

投票率較高，

which means if it wins,

意思就是，如果假資訊得逞，

disinformation can have a disproportionate impact

就可能會對選舉結果 產生不成比例的影響。

on the outcomes of elections.

So as we prepare to head into the EU elections,

因此，在歐盟選舉即將到來之前，

we created content for every EU country,

我們為每一個歐盟 成員國都創建了內容，

in 27 languages,

用了二十七種語言，

aiming to empower Europeans to spot and reject

旨在讓歐洲人有能力辨識和抵制

efforts to manipulate them before the elections.

試圖在選舉前操弄他們的各種作為。

Over the last year,

在過去一年中，

we have reached millions of people around the globe

我們用這些影片接觸到了 全世界數百萬人。

with these videos.

In Germany alone, we reached 42 million people.

光是在德國，

我們就接觸到了四千兩百萬人。

That's half the German population.

佔德國人口的一半。

And we found on average,

而我們發現，平均來說，

viewers of these videos were up to 10 percent more likely

看過這些影片的觀眾

辨識出操弄手法的可能性

to be able to identify manipulation efforts

比沒看過影片者的人高了 10%。

than those who hadn't seen those videos.

This is a winning formula.

這是個能成功的公式。

The evidence shows us

我們從證據得知，

that prebunking is effective at building resistance to disinformation.

預先闢謠可以有效協助抵制假資訊。

It begs the question,

這就帶出了一個問題，

how do we make that resistance last?

如何讓這種抵制持續下去？

How do we build long-term societal resilience

要如何建立長期的社會韌性，

to disinformation efforts?

對抗假資訊的作為？

There is an ongoing effort to use disinformation

試圖運用假資訊來破壞 我們民主的作為沒有停過。

to undermine our democracies.

Just last month,

就在上個月，美國司法部 查獲了三十二個網域，

the US Justice Department seized 32 internet domains

secretly deployed by the Russian government

皆是俄羅斯政府暗地中部署的，

to spread disinformation across the US and Europe.

目的是在美國和歐洲 各地散播假資訊。

This included deliberate efforts to exploit anxieties and fear

他們所做的包括蓄意利用民眾 對於能源轉型的焦慮和恐懼，

across the public about the energy transition,

specifically to encourage violence.

特別是要鼓勵暴力。

Now it’s not just the Russian government that we need to be worried about.

我們需要擔心的不只是俄羅斯政府。

Easy access to generative AI tools

很容易取得並使用 生成式人工智慧工具

means that anyone,

就表示任何人，

not just those with resources, money and power,

不只是擁有資源、 金錢和權力的人，

can create high-quality,

都可以創作出高品質、

effective, powerful disinformation content.

有效、強大的假資訊內容。

And the sources of disinformation are varied.

且假資訊的來源非常多。

They can come from our elected officials

可能來自我們的民選官員，

all the way through to our neighbors down the road.

一直到我們附近的鄰居都有可能。

Many of us don't need to look further than our own families.

許多人甚至會從自己的 家人那裡聽到假資訊。

But so many of the tools we tested online

但我們在線上測試的許多工具

are even more powerful

如果是直接來自 你在真實生活中（IRL）

when they come directly from the people

that you trust and love the most in real life, IRL.

最信任且最親愛的人時， 就會更強大。

So instead of endlessly debating the facts,

所以，不要無止盡地去辯論事實，

換成提供工具給你的家人， 讓他們可以在線上保護自己。

give your loved ones the tools

that they need to protect themselves online.

Information manipulation is unfortunately the new norm.

很不幸，資訊操弄是新的常態。

But that doesn't mean we need to accept our loved ones being misled,

但那並不表示我們必須 接受我們的家人被誤導，

and we shouldn't accept our climate scientists living in fear.

我們也不應該接受我們的 氣候科學家必須生活在恐懼中。

So if we can't fact-check our way out of this problem,

所以，如果我們無法靠 查證事實來解決這個問題，

we need to beat disinformation at its own game

對假資訊，我們就得 以其人之道還治其人之身，

by reaching people before disinformation does

比假資訊更早一步接觸到人，

and giving them all the tools that they need

並給予他們所需的全部工具，

to protect themselves online.

讓他們在線上保護自己。

Thank you so much.

非常謝謝。

(Applause)

（掌聲）