Name: 面對現實吧，你正在被監視 (Face It, You're Being Watched)
Uploaded: 2021-01-14T08:16:28.000Z
Duration: 3 min 54 s

In a tweet thread exchanged with JetBlue on April 17th, writer Mackenzie Fegan had a few questions about the airline's new boarding procedure.

在 4 月 17 號與 JetBlue 的推文回覆訊息中，作家 Mackenzie Fegan 對於該航空公司新的登機程序有一些疑問。

Instead, a camera and screen verified her identity against a US Customs and Border Patrol database, then let her on the plane.

取而代之的是一台相機和螢幕，在美國海關及邊境巡邏數據庫驗證她的身份過後讓她登機。

Some passengers might consider the increasing use of facial recognition in everyday life convenient, some might think it's Orwellian, but it's already everywhere.

有些乘客可能覺得在日常生活中增加臉部辨識是一種方便，其他的人則認為它是監控人民的專制主義，不過它已經無所不在了。

問題在於，這會做到什麼程度？

Turns out, even some of those developing the technology are scared of what the answer might be.

事實證明，即使是開發這項技術的人們對於答案會是什麼都感到恐懼。

This is your Bloomberg QuickTake on Facial Recognition.

這是 Bloomberg QuickTake 臉部辨識。

In May, San Francisco became the first American city to block police and other agencies from using facial recognition software.

五月，聖地牙哥成為美國第一個阻擋警察和其他機構使用臉部辨識系統的城市。

The biggest concerns are all really around civil liberties and whether you're essentially enabling a kind of totalitarian state with this technology, which seems to be sort of the way things are heading.

最令人憂心的是公民自由以及是否會因為應用這項科技而實現一個極權主義的國家，一切似乎已朝著這個方向前進。

而究竟我們是如何走到這個地步的呢？

Like other artificial intelligence applications, facial recognition initially developed slowly, beginning in the 1960s.

就像其他人工智能引用，臉部辨識開始於 1960 年代，最初發展緩慢。

With the help of newly available high-def cameras, machine learning, and giant databases of photos to increase accuracy, it advanced in a hurry.

藉由新式高解析度相機的幫助，機械學習以及龐大的相片資料庫以提高準確性，這項技術在短期內進步神速。

Facial recognition is a basic technology, it takes images from video cameras and tries to identify the faces of people in those images.

臉部辨識從攝像機拍攝照片並嘗試識別這些圖像中人物的臉部。

It does so by taking some key points in the face usually and doing measurements of the distance between all those various points.

它識別的方式是透過在臉上採取一些關鍵點並測量這些不同點之間的距離。

In December 2018, London police made their first arrest based on facial recognition, after cross-checking photos of pedestrians in tourist hotspots against a database of known criminals.

2018 年 12 月，倫敦警方利用已知罪犯數據庫，在交叉比對旅遊熱點的行人照片後，藉由臉部辨識進行首次逮捕，

In New Delhi, a police trial reportedly identified 3,000 missing children in just four days.

據說新德里一項警方的測試，在短短 4 天內就識別出 3 千名的失蹤兒童。

So if it involves catching criminals and finding missing children, why would anyone be against it?

因此如果和逮捕罪犯和搜尋失蹤兒童有關，誰會反對它呢？

For that, one might look at the most developed facial recognition network in the world, in China.

為此，人們可以看看世界上臉部辨識系統網絡最發達的國家，中國。

CCTV and facial recognition have been used in combination to create this sort of vast surveillance apparatus and that's been particularly applied against certain ethnic minority groups.

中央電視台結合臉部辨識以創造一個龐大的監視裝置，並特別運用於某些少數民族群體。

But even in western democracies there's a concern about police departments using this technology to try to find suspects, or even people who might be involved in legitimate protests, so they could be tracked.

但即使在西方民主國家，同樣有人擔心警察部門藉由這項科技找尋嫌疑犯，甚至是合法參與抗議活動的人，這樣他們能夠被追蹤。

And those are just concerns about the technology when it works as intended.

而這些僅是當臉部辨識按照計畫執行工作時所產生的憂慮.

A study from the MIT Media Lab found that white men in a sample were correctly identified 99 percent of the time, while error rates of up to 35% were found when it came to darker skinned women.

一項由麻省理工學院媒體實驗室的研究發現在樣本中的白人男性能夠 99％ 被準確識別出來，而在膚色較深的女性中，有高達 35% 的錯誤率。

It was the first major tech company to do so and say, they really feel uncomfortable with deploying this technology until there's clear regulation around it.

它是第一家站出來發言的大型科技公司，他們表示在沒有明確的規範以前，這項科技的部署讓他們感到極度地不自在。

They were then joined by Amazon, which kind of seconded those calls.

之後亞馬遜公司也加入微軟，贊成他們的呼籲。

It seems like some other companies are kind of plowing ahead without such qualms at all.

其他公司在沒有這種疑慮的情況下正向前推動。

To stake out some guidelines, the Algorithmic Justice Center and Georgetown University Law Center unveiled the Safe Face Pledge, which asks companies not to provide facial AI for autonomous weapons or sell to law enforcement unless explicit laws are debated and passed to allow it.

為了制定一些準則，演算法正義聯盟和喬治敦大學法律中心公佈安全面部誓言，要求公司切莫為自動武器提供臉部辨識人工智能或出售給執法部門，除非有明確的法律經過辯論並通過以後才能允許。

A few companies have signed on, but notably not Microsoft or Amazon, possibly loathe to lose the opportunity to sell facial recognition to police departments and governments the world over.

一些公司都已經簽約，應該是不想失去能夠販賣臉部辨識給世界各地的警察部門和政府機關的機會，但很明顯地微軟或亞馬遜公司都沒有參與。

So what's preventing your image from saying a whole lot more about you than it used to?

那麼有什麼能夠阻止你的圖像讓人能夠了解有關於你的一切，就好像從前一樣？