機(jī)器人可信嗎？

盡管電影中的智能機(jī)器人尚未成為現(xiàn)實(shí)，但隨著技術(shù)的進(jìn)步，機(jī)器人的確離我們的生活越來越近。在享受新技術(shù)帶來的便利，甚至越來越多地對(duì)機(jī)器人寄予感情的同時(shí)，人們也許是時(shí)候認(rèn)識(shí)到：不管機(jī)器人的外表被設(shè)計(jì)成什么可愛的模樣，只要它們可以被操控，可以接入網(wǎng)絡(luò)，就可能帶來危險(xiǎn)。

In February， 2015， a South Korean woman was sleeping on the floor when her robot vacuum ate her hair， forcing her to call for emergency help. It may not be the dystopian1） future that Stephen Hawking warned us about—where intelligent devices “spell the end of the human race”—but it does highlight one of the unexpected dangers of inviting robots into our home.

There are many other examples of intelligent technology gone bad， but more often than not they involve deception rather than physical danger. Malevolent2） bots3）， designed by criminals， are now ubiquitous on social media sites and elsewhere online. The mobile dating app Tinder， for example， has been frequently infiltrated by bots posing as real people that attempt to manipulate users into using their webcams4） or disclosing credit card information. So it’s not a stretch5） to imagine that untrustworthy bots may soon come to the physical world.

Meanwhile， increasing evidence suggests that we are susceptible6） to telling our deepest， darkest secrets to anthropomorphic7） robots whose cute faces may hide exploitative code—children particularly so. So how do we protect ourselves from double-crossing8） deceptions？

Once you’ve invited a bot into your home， you need to manage your expectations. Movies and marketing may have primed9） us to expect sophisticated interaction with our robotic chums10） but we’ve still got a long way to go before they are as socially aware as they are often depicted. Given the gulf between expectation and reality， it’s important to avoid being tricked by a fake-out known as a “Wizard-of-Oz setup11），” where users are led to believe that robots are acting autonomously when， in fact， human operators are remotely controlling some of their operations.

Misjudging where behaviour originates can be an especially acute problem in cases where consumers feel so comfortable with a non-sentient12） device that they reveal intimate information that they would have withheld had they known a human was in the loop13）. Take the service “Invisible Boyfriend” for example. For a monthly subscription， romantic texts and voicemails are sent to your phone from a faux lover. Although the company initially sought to make the fake beau14） fully automated， the technology wasn’t sophisticated enough， so in reality， human workers generate the amorous15） exchanges. But not all customers understand how the system works， and thanks to the hype16） surrounding artificial intelligence and well-documented cases of automated bots successfully tricking people into believing that they’re real humans， some people might erroneously17） believe they’re receiving computer-composed dialogue.

The take-home message is clear： As robots become increasingly connected to the internet， and able to respond to natural language， you need to be especially vigilant18） about figuring out who or what you’re talking to.

We also need to think long and hard about how information is being stored and shared when it comes to robots that can record our every move. Some recording devices may have been designed for entertainment but can easily be adapted for more nefarious19） purposes. Take Nixie， the wearable camera that can fly off your wrist at a moment’s notice and take aerial shots around you. It doesn’t take much imagination to see how such technology could be abused.

Most people guard their secrets in the presence of a recording device. But what happens once we get used to a robot around the house， answering our every beck20） and call？ We may be at risk of letting our guard down， treating them as extended members of the family. If the technology around us is able to record and process speech， images and movement—never mind eavesdrop21） on our juiciest secrets—what will happen to that information？ Where will it be stored， who will have access？ If our internet history is anything to go by22）， these details could be worth their weight in gold23） to advertising companies. If we grow accustomed to having trusted robots integrated into our daily lives， our words and deeds could easily become overly-exposed.

So， what is the safest way to welcome robots into our homes， public spaces， and social lives？ We should be cautiously optimistic that these intelligent machines could become enriching companions， while acknowledging that we need to determine strict boundaries for robots capable of deception and manipulation.

We might think of expanding the reach of consumer protection agencies or creating new robotic-centric policies. Just as the advent24） of radio called for the formation of the Federal Radio Commission in the US， advances in robotics may call for the advent of a body responsible for the integration of robotics into society. Someone to turn to should your robot commit a crime， steal your credit card … or try to eat your hair.

2015年2月，一名韓國(guó)女子在地鋪上睡覺時(shí)，掃地機(jī)器人“吃”了她的頭發(fā)，迫使她撥打緊急求救電話。這或許并不是斯蒂芬·霍金警告過我們的反烏托邦式的未來—在那樣的未來中，智能裝置“預(yù)示著人類的末日”—但這件事確實(shí)凸顯了把機(jī)器人請(qǐng)進(jìn)家門所帶來的意想不到的危險(xiǎn)之一。

智能技術(shù)出問題的其他例子還有很多，但這些例子涉及的往往是欺詐，而非人身危險(xiǎn)。如今，在社交媒體網(wǎng)站及其他網(wǎng)站上，犯罪分子設(shè)計(jì)的惡意機(jī)器人程序無處不在。例如，手機(jī)約會(huì)應(yīng)用軟件Tinder就經(jīng)常被機(jī)器人程序滲透，這些程序冒充真人，試圖左右用戶，讓他們使用網(wǎng)絡(luò)攝像頭或者透露信用卡信息。因此，不可靠的機(jī)器人或許很快就要來到現(xiàn)實(shí)世界，這種猜想并不夸張。

同時(shí)，越來越多的證據(jù)表明，我們—尤其是兒童—很容易把自己內(nèi)心最深處、最隱秘的秘密透露給人形機(jī)器人，它們可愛的面容背后可能隱藏著利用程序漏洞的代碼。那么我們?cè)撊绾伪Ｗo(hù)自己免遭背叛性的欺騙呢？

一旦你把一個(gè)機(jī)器人請(qǐng)入家中，你需要控制自己的預(yù)期。電影和營(yíng)銷可能事先讓我們有所期待，期待與我們的機(jī)器人朋友進(jìn)行復(fù)雜的交流，但是在機(jī)器人像人們經(jīng)常描述的那樣具備社交意識(shí)之前，我們還有很長(zhǎng)的路要走。鑒于預(yù)期與現(xiàn)實(shí)之間的鴻溝，要避免陷入被稱作“奧茲國(guó)魔法師圈套”的騙局，這很重要。這種騙局誘導(dǎo)用戶相信機(jī)器人是自主行事的，而事實(shí)上真人操作者正在遙控機(jī)器人的某些活動(dòng)。

跟一個(gè)沒有感知力的裝置在一起，消費(fèi)者感到非常輕松自在，于是就會(huì)透露私密的信息。要是他們知道背后有一個(gè)真人參與其中，是不會(huì)透露這些信息的。在這些情況下，對(duì)行為來源的錯(cuò)誤判斷就會(huì)成為尤其嚴(yán)重的問題。以“隱形男友”這項(xiàng)服務(wù)為例。如果你訂購(gòu)月度服務(wù)，就會(huì)有一位虛擬情人向你的手機(jī)發(fā)送浪漫的信息和語音郵件。盡管推出該項(xiàng)服務(wù)的公司最初企圖讓假情郎完全自動(dòng)化，但技術(shù)還不夠成熟，因此那些愛意綿綿的交流內(nèi)容其實(shí)是由真人員工發(fā)出的。然而，并非所有消費(fèi)者都明白該系統(tǒng)的運(yùn)作方式，再加上人工智能被大肆炒作，且有大量記錄的案例顯示自動(dòng)化的機(jī)器人程序成功地讓人們錯(cuò)以為它們是真人，因而有些人可能會(huì)誤以為自己收到的是計(jì)算機(jī)生成的對(duì)話。

從中獲得的關(guān)鍵信息顯而易見：隨著機(jī)器人越來越多地接入互聯(lián)網(wǎng)，而且能夠回應(yīng)自然語言，你需要特別警惕，搞清楚你正在和誰或什么交談。

對(duì)于能夠記錄我們一舉一動(dòng)的機(jī)器人，我們還需要認(rèn)真思考信息是如何被存儲(chǔ)和分享的。有些錄制設(shè)備的設(shè)計(jì)初衷可能是娛樂，但稍加改造就能被用于更加邪惡的目的。比如可穿戴攝像機(jī)Nixie，它一收到指示就可以立即從你的手腕上起飛，在你的四周進(jìn)行空中拍照。不難想象這樣的技術(shù)會(huì)如何被濫用。

身邊有錄制設(shè)備時(shí)，大多數(shù)人都會(huì)保守自己的秘密。可是，一旦我們習(xí)慣了有一個(gè)機(jī)器人在家中而且任憑我們使喚，又會(huì)發(fā)生什么呢？我們可能面臨放松警惕的危險(xiǎn)，將機(jī)器人視為大家庭中的成員。如果我們身邊的技術(shù)能夠記錄并處理語言、圖像和動(dòng)作—更不用說竊聽我們最富刺激性的隱私—這些信息會(huì)有怎樣的遭遇？信息會(huì)被儲(chǔ)存在哪里，誰能夠獲取這些信息？如果我們根據(jù)互聯(lián)網(wǎng)的歷史來判斷就會(huì)明白，這些詳細(xì)信息對(duì)于廣告公司來說會(huì)非常有價(jià)值。如果我們逐漸習(xí)慣了讓被信賴的機(jī)器人融入日常生活，我們的一言一行就很容易遭到過度暴露。

那么，要迎接機(jī)器人進(jìn)入我們的家中、公共場(chǎng)所和社會(huì)生活中，怎樣才是最安全的方式？對(duì)于這些智能的機(jī)器能否成為豐富我們生活的伴侶，我們應(yīng)抱著謹(jǐn)慎樂觀的態(tài)度，同時(shí)也要承認(rèn)，我們需要針對(duì)能夠進(jìn)行欺詐和操控的機(jī)器人劃定嚴(yán)格的界限。

我們可以考慮擴(kuò)大消費(fèi)者保護(hù)機(jī)構(gòu)的職權(quán)范圍，或者制定針對(duì)機(jī)器人的新政策。正如無線電的誕生需要美國(guó)成立聯(lián)邦無線電委員會(huì)一樣，機(jī)器人技術(shù)的發(fā)展可能也需要成立一個(gè)機(jī)構(gòu)，為機(jī)器人技術(shù)融入社會(huì)負(fù)責(zé)。如果你的機(jī)器人犯下罪行，偷了你的信用卡……或者試圖吃掉你的頭發(fā)，得有個(gè)可以求助的人。

1.dystopian [d?s?t??pi?n] adj. 反面烏托邦的，反面假想國(guó)的

2.malevolent [m??lev?l?nt] adj. 含有惡意的；惡毒的

3.bot [b?t] n. 機(jī)器人程序；（科幻小說中的）機(jī)器人

4.webcam [?web?k?m] n. 網(wǎng)絡(luò)攝像頭

5.stretch [stret?] n. 夸張

6.susceptible [s??sept?b（?）l] adj. 易受影響的

7.anthropomorphic [??nθr?p??m??（r）f?k] adj. 被賦予人形（或人性）的，擬人的

8.double-cross：出賣；欺騙

9.prime [pra?m] vt. 事先為……提供情況（或消息、情報(bào)、事實(shí)等）

10.chum [t??m] n.〈口〉好友

11.Wizard-of-Oz setup：此處指人機(jī)交互領(lǐng)域的一種實(shí)驗(yàn)方法。實(shí)驗(yàn)過程中，受試者與電腦系統(tǒng)進(jìn)行互動(dòng)，實(shí)際上電腦由背后的真人操控或部分操控。

12.sentient [?sen?（?）nt] adj. 有感覺力的，有知覺力的

13.in the loop：在決策圈內(nèi)；在消息圈內(nèi)

14.beau [b??] n.〈主美〉男友，情郎；（向女子）獻(xiàn)殷勤的男人

15.amorous [??m?r?s] adj. 求愛的；表示愛情的

16.hype [ha?p] n. （刺激興趣、促進(jìn)銷售等的）天花亂墜的廣告宣傳

17.erroneously [??r??ni?sli] adv. 錯(cuò)誤地，不正確地

18.vigilant [?v?d??l?nt] adj. 警戒的；警覺的

19.nefarious [n??fe?ri?s] adj. 惡毒的，邪惡的

20.beck [bek] n. （以點(diǎn)頭或手勢(shì)表示的）召喚；命令，吩咐

21.eavesdrop [?i?vz?dr?p] vi. 偷聽，竊聽

22.go by：根據(jù)……做出判斷

23.worth one’s weight in gold：非常有價(jià)值的

24.advent [??dvent] n. （尤指不尋常的人或事物的）出現(xiàn)，到來