Name: 智能音響很酷炫沒錯！小心你的個資被偷聽 (The real cost of smart speakers)
Uploaded: 2022-02-14T21:01:20.000Z
Duration: 5 min 35 s

In 2013, Google announced it was adding something new to its Chrome browser:

2013 年，Google 宣布期將會為下的 Chrome 瀏覽器增添個新功能：

Ok, Google. Show me things to do in Santa Cruz.

OK，Google。告訴我可以在聖塔·克魯茲做的事情。

Here are popular attractions in Santa Cruz.

以下是聖塔·克魯茲的熱門景點。

Voice recognition that you could activate by speaking a keyword.

讓你對著鍵盤講話就能啟動的語音識別功能。

In 2014, Apple updated its Siri function to work in the same way.

2014 年，蘋果更新了 Siri 功能，讓其能以同樣的方式運作。

嘿，Siri，秀一些義大利餛飩的圖片給我看。

以下是幾張義大利餛飩的圖片。

Voice activation was a flashy new addition to these companies' existing browsers and phones.

語音聲控成為了這些公司旗下的瀏覽器與手機的全新酷炫功能。

But then, a few months later, Amazon released something much simpler:

但是，幾個月後，Amazon 發佈了更簡單明瞭的產品：

When it first arrived from Amazon, I didn't know what it was.

Amazon 把這個東西送過來的時候，我還搞不懂這到底是什麼。

A product whose sole purpose was to interact with your voice.

只透過你的聲音即可互動的產品。

And you can see in this video they made, that they had to convince people that it would be useful.

而你可以從他們製作的這個影片中看出來，他們當時得說服人們這是實用性高的產品。

I can play music, answer questions, get the news and weather, create to-do lists and much more.

我可以播放音樂、回答問題、查詢新聞與天氣、建立代辦清單等更多用途。

But it worked. Today, every 1 in 5 American households has a smart speaker.

但它確實辦到了。如今，每 5 個美國家庭中就有 1 個智慧音響。

Most are made by Amazon, Google, and Apple, and their technology doesn't just sit on tabletops anymore.

其中大多數是 Amazon、Google 和 Apple 製造的，且他們的科技再也不僅僅侷限於擺放在桌面上。

You can find Alexa in your ears; around your wrist, or just a finger.

你可在耳朵裡、手腕，或甚至是手指上連結上 Alexa。

She's even right in front of your eyeballs.

You can also find smart-speakers built into remote controls, security cameras, and doorbells.

你也能在遙控器、監視器和門鈴上找到內建的智慧音響。

Or in the form of an adorable tiger, so that it's easier for your kids to use.

或者被做成可愛的老虎形狀，讓你的孩子們可以更輕易地使用。

And each Alexa product can reportedly process 100,000 different commands.

而據說每個 Alexa 可以處理 10 萬個不同的指令。

But all this convenience does come at the expense of some of our privacy.

但這一切的便利都是以犧牲我們的一些隱私做為代價的。

So it's worth asking: What are we really giving up when we use one?

所以值得一問的是：在我們使用這些產品的同時，到底放棄了什麼？

The first thing to understand is that smart speakers aren't actually that smart.

首先要明白的是，智慧音響其實並沒有那麼有智慧。

It probably seems maybe like there's a lot more going on than there is.

乍看之下或許感覺有很多事情在背後發生。

Sara Morrison covers personal data and privacy for Recode.

Sara Morrison 是為 Recode 網站報導個資和隱私相關議題的記者。

But essentially, um, it's a microphone  that connects to the Internet.

但本質上，它其實就是一個連接上網路的麥克風而已。

When I ask it about the weather—"Alexa, what's the temperature?"—It's not the thing that finds the temperature.

當我問它天氣的時候－「Alexa，現在氣溫幾度？」－其實並不是眼前的音響在測量溫度。

It just records my question, sends that recording to an Amazon server, which analyzes it and finds the temperature.

它只是把我的問題錄下來，把這個錄音發送到 Amazon 的伺服器上後，再由伺服器分析錄音並查詢溫度。

What seems like a genius digital helper, is really just a direct line to Amazon.

看似天才的數字幫手，其實只是條直達 Amazon 的熱線而已。

And while I soon forget my question, Amazon remembers it.

而雖然我很快就忘記了自己問了什麼，Amazon 卻還會記得。

In my speaker's settings, I can see that my recordings are actually saved, until I delete them. So….

在我的音響設置中，我可以看到我的錄音居然被儲存了下來，直到我手動刪除它們為止。所以....

Unless you set it somehow differently, it will pretty much stay on their servers indefinitely.

除非你以某種不同的方式調整設定，否則它幾乎會無限期地留存在他們的伺服器上。

According to a recent survey, almost half of smart-speaker owners incorrectly believed their audio was only saved temporarily or not at all.

根據最近的一項調查顯示，近半數的智慧音響用戶錯誤地認為他們的語音錄音只會被暫時儲存，或根本沒有被儲存下來。

而 47% 的人認為這是不可接受的。

And then there's the question of what they do with those recordings.

另一個問題則是，他們會如何處理這些錄音。

還會有其他人聽到我的問題嗎？

I think most people don't really know that they do send that recording to a server, that humans might listen to your recording,

我想大多數人並不知道他們確實把那段錄音發送到了伺服器上，而如此一來便會有其他人可能會聽到你的錄音，

which sounds kind of creepy, but it's really for accuracy purposes, to make sure that they're responding to the right things.

這聽起來有點嚇人， 但其實這只是為了語音辨識的準確目的，確保它們的回覆是正確的。

But obviously, if you don't know that that's a possibility that can happen, you might not really like that.

但很顯然的，如果你不知道有這種可能性，你可能並不會喜歡這件事。

事實上，31% 的人不喜歡這種做法。

We give them a lot of this data, we expose a ton of our information to them, and I mean, I don't know, ultimately, what they're using it for, and neither do you.

我們給了他們大量的數據，並把大量的資訊暴露給他們。我不知道最終他們使用這些資料的目的，而你也不知道。

One thing we can assume they're using our voice data for is some form of ad targeting.

我們可以假設他們其中一個目的，是使用我們的語音資料來進行某種形式的廣告投放。

Google says it doesn't use the recordings themselves, but may use a text read-out of your recording to show you personalized ads.

Google 表示他們不會使用錄音本身，但可能會使用錄音分析出的文本來投放個人化的廣告。

Amazon has been more vague. It says it doesn't use your data to target ads, but that it may use it on Amazon.com, which of course is powered not by ads, but by sophisticated product recommendations.

Amazon 的說法則比較模糊。它說它不會使用你的資料來投放廣告，但可能會在 Amazon 網站上使用這些資料，而這指的當然就是其複雜的產品推薦系統。

What's clear, though, is that they feel very confident about how valuable our voice data could be.

但很明顯的是，他們非常確信我們的語音資料能造就相當的價值。

That's why they're adding smart-speaker technology to so many different products, so that they can collect more.

這就是他們會在這麼多不同的產品中加入智慧音響技術的原因，以便他們能收集更多語音資料。

It's also partly why smart-speakers are so cheap; as little as 10 bucks in the US, because the real price is our data.

這也是智慧音響如此便宜的部分原因。在美國一個最低只要 10 美元，而背後原因是因為真正的價格其實是我們的資料本身。

There's a longer-term thing that some privacy experts have sort of brought up to me,

一些隱私專家向我提出了一個更加長遠的看法，

which is, you know, the more common they are, the more comfortable we get with them in every aspect of our lives,

也就是我們越常看到這些產品，越習慣它們存在我們生活的各方面後，

the more comfortable we kind of maybe get with there being things that take data, and information exposure all around our lives.

我們可能也會對在我們生活周遭收集數據與資料的行為開始感到習慣。

In 2019, Amazon announced its new "Sidewalk" feature.

在 2019 年，Amazon 發表了新的 Sidewalk 功能。

When implemented, it will connect your smart speaker with others in your vicinity, like the ones in your neighbor's homes.

開啟這個功能後，它會把你的智慧音響與家裡附近，例如鄰居家裡的智慧音響一起串聯起來。

But you wouldn't know that unless you checked your settings.

但除非你有檢查過裝置的設定，不然你根本不會察覺。

這是需要主動選擇退出，而不是選擇加入的功能。

When something's opt-out, it means that they want as many people as possible to sort of be doing it.

當某個功能是「選擇退出」型的，代表他們想要盡可能有更多人使用這個功能。

Thanks to public pressure, tech companies have made it easier to opt-out of some data-sharing features.

迫於大眾的壓力，科技公司現在把選擇退出某些資料分享功能的手續變得更加簡單。

On Amazon products, you can set your recordings to automatically delete, and prevent Amazon from using your recordings to train its system.

在 Amazon 的產品上，你可以調整系統將你的錄音自動刪除，並防止 Amazon 使用你的錄音來訓練其系統。

Google has an option to turn off ad personalizations.

Google 則有一個能夠關閉個人化廣告的選項。

The reason that they have done that is because they've kinda gotten in trouble. Or people have found them doing things that they didn't know about and got upset.

他們之所以這樣做是因為他們已經有點惹上麻煩了。或者人們發現了他們在做自己以前不知道的事情，所以為此感到憤怒。

Always makes you wonder, what is the thing we're going to catch them doing later that they're doing right now?

而這總會讓你懷疑，我們以後還會抓到他們現在搞不好還在秘密進行的哪些事情？

The huge popularity of smart-speakers shows that we've already entered a new wave of technology, made by companies whose goal is to learn everything they can about us.

智慧音響的大行其道說明了我們已經邁入了一個新的科技浪潮，而這樣的趨勢是由那些想要了解我們生活中一切事物的公司所造成的。

And what they might do with that information is still a mystery.

而他們還可能拿這些資訊來做什麼，仍舊是個謎。