史上最奇怪的硬體產品：谷歌新型智能剪輯相機

來自專欄 外刊搬運工

每日知乎專欄與微信公眾號同步更新，公眾號：外刊搬運工，ID：ForeignMagzine.

Google』s new camera decides what to photograph, based on AI algorithms

根據AI演算法自動決定按快門

原文：David Pogue，《科學美國人》2018年5月刊

翻譯：鳳梨是只胖柯基

校閱：PC

Google』s Clips camera is a tiny sliver of a camera, the size of two Wheat Thins crackers. You can set it down anywhere or clip it to anything. Once you turn it on, you don』t have to press a button or use a self-timer to take pictures. The camera decides when to snap, based on Google』s artificial-intelligence algorithms.

谷歌新型智能剪輯相機（又稱「夾子相機」）是一種微型相機，大小約為兩片小麥脆餅。你可以把它放在任何地方，或者把它夾在任何地方。一旦你打開它，你就不必按下拍照按鈕或使用自拍器拍照。根據谷歌的人工智慧演算法，相機會自己決定何時拍攝。

The Clips』s heart is in the right place. It solves some real problems for its target audience, which is parents (of kids or of pets).

夾子相機的定位很正確。它解決了目標受眾群體的一些實際問題，即父母(想要拍孩子或寵物)。

First, if you』re in that category, you』re probably never in any of your own photographs, because you』re always behind the camera.

首先，如果你養了孩子或者寵物，你可能永遠不會在你自己的照片里，因為你總是在相機後面拍照。

Second, babies and young children often stop whatever cute thing they』re doing the moment you pull out your phone. They get distracted by it or feel self-conscious. But the Clips avoids that problem because it』s unobtrusive and because you』re not holding it between your face and the kid』s.

其次，當嬰兒和小孩正在做可愛的事情時，你一拿出手機想記錄這一刻，他們卻因你拿手機這個動作被它分散注意力或感到難為情。但是「夾子相機」避免了這個問題，因為它是不引人注目的，你也不用舉著它對準你的或者孩子的臉。

Truth is, I suspect the Clips will probably flop. The camera isn』t very impressive next to those in some smartphones, and $250 is a steep price for a one-trick pony. But its central idea—

AI as photographer—is fascinating.

事實是，我懷疑這些剪輯可能並不好用。有了智能手機，攝像頭並不是很有吸引力。250美元的價格買一個小玩意還是很昂貴的。但它的理念——人工智慧作為攝影師還是很贊的。

AI isn』t organic. It has to be programmed —taught or coded by engineers. In other words, the AI doesn』t ultimately decide what makes a good picture; its programmers, informed by photography experts, do.

人工智慧沒有思想，它必須由工程師編程或編碼。換句話說，人工智慧並不能最終決定什麼是好的畫面。它的程序員從攝影專家們聽取建議什麼時候應該按快門。

Some of the AI』s decision making in the Clips is obvious. It looks for scenes of activity. It favors familiar subjects—people whose faces it sees most often. It avoids capturing an image when something is blocking the lens, like your fingers or your grabby baby』s hands. It prefers good lighting. It takes its best shots three to eight feet away.

AI在「夾子相機」中的決策是顯而易見的，它喜歡活動的場景，也更傾向於熟悉的物品，熟悉的人。當某些東西擋住了鏡頭時，比如你的手指或被抓住的寶寶的手，它就不會捕捉圖像。它也喜歡很好的照明環境，最好把它放在三到八英尺遠的地方拍攝。

But here』s where things get more complicated: The camera is also designed to wait for happy facial expressions. It tends not to capture anybody who is sad, angry, sleepy, bored or crying.

更複雜的是:相機也被設計用來等待快樂的面部表情。它往往不會拍一個悲傷、憤怒、睏倦、無聊或哭泣的人。

That AI rule, unfortunately, rules out a lot of great picture taking. Let』s face it—a young child』s life is full of micro tragicomic moments that might be worth recording, even if they produce brief bursts of tears. You know: His ice cream falls off the cone onto the floor. A puppy licks her face a little too energetically. A well-meaning clown scares him.

不幸的是，人工智慧規則其實排除了大量的圖片。讓我們面對現實——一個小孩子的生活充滿了微小的悲劇性瞬間，即使他們會流淚，但也值得紀念。比如：他的冰淇淋掉到地板上了；一隻小狗太用力地舔她的臉；又或者是一個善意的小丑把他嚇了一跳。

Google is aware of the problem and plans to add a new preference setting—not a check box called 「Include Misery」 but an option that makes the camera watch for changes in facial expression.

谷歌意識到了這個問題，並計劃增加一個新的偏好設置——不是添加一個叫做「包含悲傷」的複選框，而是新增一個讓攝像頭觀察面部表情變化的選項。

In the meantime, the Clips』s preference for joyous moments tends to exaggerate two happiness filters we already put on our lives.

與此同時，「夾子相機」對快樂時刻的偏愛，往往誇大了我們已經在生活中已經使用了的兩個幸福過濾器。

First, we already self-edit our video and photographic memories simply by choosing what to shoot. Most people, most of the time, record high points such as celebrations and travel. Your collection probably contains very few pics of you fighting with your spouse, depressed by your job or in pain from an injury.

第一個是，我們已經通過選擇拍攝什麼來自行編輯我們的視頻和照片記憶。大多數人，大部分時間，都記錄著諸如慶祝和旅行這樣的歡樂時光。你的照片收藏里可能只包含了幾張你和你的配偶吵架，你的工作煩惱或受傷的痛苦的照片。

Second, we further curate our recordings by choosing which to post online. At this point, we don』t just risk deceiving ourselves about the overall happiness balance in our lives; we』re explicitly trying to paint a picture of a wonderful life for our followers. We become brand ambassadors for our supposedly flawless lives.

第二個是，我們通過選擇在網上發布的方式來進一步策劃我們的生活軌跡。在這一點上，我們不只是在冒險欺騙自己的生活中整體的幸福平衡;我們明確地試圖為我們的社交媒體關注者們描繪一幅美好生活的畫面。我們成為了我們所謂的完美生活的品牌大使。

Studies have shown that the result of all this happy filtering can sadden other people on social media, who develop 「Facebook envy.」 You begin to wonder why we take pictures and videos in the first place. What』s the purpose of those acts? Is it to create a faithful record of our lives, both high and low moments? Is there anything wrong with immortalizing only the bright spots, permitting the darker stuff to fade out of view—and maybe out of memory?

研究表明，所有這些快樂過濾的結果會讓社交媒體上的其他人感到悲傷，甚至會發展出「臉書型嫉妒」。「你開始懷疑起我們拍照片和視頻的初衷是什麼。」這些行為的目的是什麼?它是否為我們創造了生活的忠實記錄，不管是高興還是低潮的時候?只保留生命中的閃光點，讓黑暗的部分淡出視線——甚至淡出記憶，是否有什麼不對?

Answering those questions depends, in part, on who your audience is. An older you? Your descendants? Your Facebook friends? There』s no right answer. We all take and curate pictures—or don』t—for different reasons. If Google』s Clips camera achieves nothing more than throwing those questions into sharper focus, its invention won』t have been in vain.

回答這些問題，部分取決於誰是你的觀眾。一個年長你的人?你的後代嗎?你的臉書朋友嗎?這沒有正確的答案。我們都是出於不同的原因去拍攝視頻或照片。如果谷歌的「夾子」讓這些問題得到了關注，那麼它的發明將不會是徒勞的。

允許轉載，轉載需標明出處及作者、翻譯、校閱。

每日知乎專欄與微信公眾號同步更新，公眾號：外刊搬運工，ID：ForeignMagzine.