“Be aware of how the balance of control is shifting”

“注意控制的平衡是如何变化的”

NB: This article was first published in 2012 and was copied here in 2020.

注意：本文于2012年首次发表，并于2020年在此处复制。

For as long as humans have been inventing new technologies, we have tried to use these technologies to create automatons in our image. In other words, robots that we can talk to.

只要人类一直在发明新技术，我们就试图使用这些技术在我们的形象中创造自动机。 换句话说，我们可以交谈的机器人。

Our oldest myths include robots made from clay. With the advent of metalworking, Hephaestus, blacksmith to the ancient greek gods, built handmaidens out of gold. With the spread of mechanics, René I-think-therefore-I-am Descartes is said to have built a daughter out of metal and sprockets. Following the industrial revolution, steam and then electrical robots proliferated including a (fictional) Thomas-Edison-like inventor introducing the term “android” in The Future Eve.

我们最古老的神话包括用粘土制成的机器人。 随着金属加工的出现，赫菲斯托斯(Hephaestus)，古代希腊神的铁匠用黄金建造了女仆。 随着机械学的传播，据说勒内·我(RenéI-Think-I-am Descartes)用金属和链轮打造了一个女儿。 在工业革命之后，蒸汽机和电动机器人激增，其中包括一个(虚构的)托马斯·爱迪生式发明家，在“未来平安夜 ”中引入了“机器人”一词。

People have been predicting that talking robots are just 5 years away … for about 50 years now. Our expectations have exceeded the complexity that is necessary to make this a reality. A good example is from 2001: A Space Odyssey:

人们一直在预测，说话的机器人距离现在只有5年……大约50年了。 我们的期望超出了实现这一目标所必需的复杂性。 一个很好的例子是2001年的《太空漫游》 ：

“I’m sorry, Frank, I think you missed it … “

“对不起，弗兰克，我想你错过了……”

In 1968 — a year before the moon landing — audiences balked at the idea that a computer would be competitive with a human in chess by 2001, but no one batted an eyelid that the computer was carrying on a conversation in English at the same time. Not only was the English flawless, the computer had mastered the art of subtle condescension. The condescension was more than linguistic expertise — it signaled that perhaps the robot was not the willing servant and that is possessed a hidden agenda that gave it more power than its human counterparts realized.

1968年，也就是月亮登月的前一年，听众对计算机在2001年之前可以与国际象棋竞争的想法感到震惊，但没人敢说，该计算机同时在进行英语对话。 英语不仅完美无缺，而且计算机还掌握了微妙的屈尊的艺术。 谦逊不仅仅是语言专业知识-它表明机器人可能不是自愿的仆人，并且拥有隐藏的议程，赋予了它比人类同行所意识到的更大的力量。

What was the name of the robot in 2001: A Space Odyssey? To test your knowledge, answers for this robot (#1) and 19 others below are at the end of the article.

2001年该机器人的名字是：太空漫游。 为了测试您的知识，本文结尾处提供了该机器人(＃1)和下面其他19个机器人的答案。

Aristotle suggested in Politics (322 BCE) that automatons could bring the abolition of slavery. The term “robot” didn’t come until relatively recently in Čapek’s Rossum’s Universal Robots, published in 1920. Echoing Aristotle, “robot” is derived from the Czech word “rab”, which means “slave”. The tension has remained: are robots our servants and should we fear that they will rise up to control us?

亚里士多德在《 政治学》 (公元前322年)中提出，自动机可以废除奴隶制。 直到1920年Čapek的Rossum的Universal Robots中才出现“机器人”一词。呼应亚里斯多德的说法，“机器人”源自捷克语“ rab”，意思是“奴隶”。 紧张仍然存在：机器人是我们的仆人，我们是否应该担心机器人会崛起来控制我们？

So where is your talking robot? And should you be worried about it rising up to control you? You may not be having a conversation with a robot every day, but you are interacting with artificial intelligence which is understanding your communications and making decisions that are changing your life in profound ways. They cannot produce flawless sentences, but they can understand language at a scale beyond that of any one person. These robots are working with you to make you smarter, choose your friends, select your news, protect your health, move you around, and locate your enemies.

那么您的说话机器人在哪里？ 您是否应该担心它会上升来控制您？ 您可能并非每天都在与机器人对话，但是您正在与人工智能进行交互，后者正在理解您的通信并做出可以深刻改变您生活的决策。 他们不能说出完美的句子，但是他们可以理解的语言范围超出任何人。 这些机器人与您合作，可以使您变得更聪明，选择朋友，选择新闻，保护您的健康，四处走动并找到敌人。

We’ll go through examples of these influences one by one:

我们将一一列举这些影响的示例：

使您更聪明 (Making you smarter)

In the US, the Graduate Record Examination (GRE) is taken by every student applying to graduate school. One of the examiners for the written task is an automated process, grading the quality of your writing by evaluating a complex combination of the words that you use and the structure of your sentences.

在美国，每位申请研究生院的学生都参加研究生成绩考试(GRE)。 书面任务的考官之一是自动化过程，它通过评估您使用的单词和句子结构的复杂组合来评定您的写作质量。

The robot grading your GRE may not be speaking back to you directly, and there are still human graders for every exam. It is not the robots alone who have determined a million or more educational futures, but it is likely that the robots have been the deciding vote in thousands of borderline cases.

为GRE评分的机器人可能不会直接与您对话，并且每次考试仍然有人工评分。 决定一百万或更多的教育未来的不是机器人，而是机器人在成千上万起边缘事件中起了决定性的作用。

选择你的朋友 (Choosing your friends)

Every social network employs some form of language technology. The understanding of your communications is one of the factors that determine what appears in your social news feed or trending topics.

每个社交网络都采用某种形式的语言技术。 对您的交流的理解是确定出现在社交新闻源或趋势主题中的因素之一。

For some people — an increasing number of the most online — this will mean that the majority of their interactions with friends and family are being mediated by robots understanding their communications and making decisions about what to show and what to leave out.

对于某些人-在线人数最多的人越来越多-这意味着，他们与朋友和家人的大部分互动都是由机器人来理解的，这些机器人了解他们的交流方式并做出有关展示和遗忘的决定。

过滤您的新闻 (Filtering your news)

If you also get your news from social media, then the same bots that determine your news feed are also determining the news that you consume. If read the news online, then chances are there are also some language technologies that are dynamically selecting what appears to be read, as online news services and aggregators are increasingly offering personalized services by observing your reading habits and guessing your preferences.

如果您还从社交媒体获取新闻，那么确定新闻源的机器人也将确定您消费的新闻。 如果在线阅读新闻，那么随着在线新闻服务和聚合商通过观察您的阅读习惯并猜测自己的喜好，越来越多地提供个性化服务，因此还有一些语言技术可以动态选择看似阅读的内容。

And if you only watch broadcast news? Pretty much every news agency now uses social media to decide what to broadcast, so while the actual broadcast might be by humans, there were robots helping to select each story along the way.

如果只看广播新闻？ 现在几乎每个新闻社都使用社交媒体来决定广播的内容，因此，尽管实际广播可能是人为进行的，但沿途都有机器人帮助选择每个故事。

保护您的健康 (Protecting your health)

Disease outbreaks are among the world’s greatest killers. A well-known case being the 1918 “Spanish Flu” that killed more people than World War I. The first signs of a new outbreak are often local reports in plain language. In 2009, an obscure report about a “strange epidemic outbreak” in La Gloria, Mexico (pop 3,000) was not noticed by human epidemiologists first, but by GPHIN, a system in Canada that reads information about potential outbreaks and alerts health professionals. This outbreak became known as H1N1, or more simply Swine Flu — a robot got there first.

疾病暴发是世界上最大的杀手之一。 一个著名的例子是1918年的“西班牙流感”，它造成的死亡人数比第一次世界大战还多。新爆发的最初迹象通常是当地人以朴素的语言报道。 2009年，人类流行病学家首先没有注意到关于墨西哥拉格洛里亚市“奇怪的流行病暴发”(流行3,000)的晦涩报告，而是加拿大的GPHIN系统，该系统读取有关潜在暴发的信息并向卫生专业人员发出警报。 这次爆发被称为H1N1 ，或更简单的说是猪流感-机器人首先到达那里。

Health is one area where the machines are very quickly gaining on humans. We may only be making incremental progress in machines that understand human communication, but it is out-pacing the rate at which we are increasing the ratio of doctors to patients.

健康是机器在人类身上Swift发展的领域之一。 我们可能只会在理解人际交流的机器上取得进步，但是它已经超过了我们增加医生与患者比率的速度。

移动你 (Moving you around)

So what will be the first talking robot that you own? There is a good chance that it will be a self-driving car with a voice-recognition system.

那么，您拥有的第一个说话机器人将是什么？ 它很有可能会成为带有语音识别系统的自动驾驶汽车。

In other words, your first robot will be one that you sit inside, the doors will lock, and it will take you down a busy road at high speed. That’s a lot of trust.

换句话说，您的第一个机器人将是您坐在里面的机器人，门将被锁定，它将带您以高速驶向繁忙的道路。 那是很多信任。

找到你的敌人 (Locating your enemies)

The 9/11 masterminds Khalid Sheikh Mohammed and Ramzi bin al-Shibh became upset at the lack of attribution they were receiving, and so in 2002 they allowed an al-Jazeera reporter to record an audio interview bragging about their involvement. The reporter was blindfolded and taken to an unknown location, but within days of airing voice-recognition technology from the interview is believed to have been used to pinpoint their location from satellite phones, leading to their arrests.

9/11策划人Khalid Sheikh Mohammed和Ramzi bin al-Shibh由于缺乏归因而感到沮丧，因此在2002年，他们允许半岛电视台的一名记者吹嘘他们的参与。 记者被蒙住了眼睛，被带到一个未知的地方，但是据传，采访中的语音识别技术在几天之内就被用于从卫星电话中查明他们的位置，从而导致他们被捕。

It is one version of the future that we were promised in cartoons: evil geniuses are doing battle with artificial intelligences that control satellites, ultimately being brought down by their own egos.

这是我们在动画片中所应许的未来的一个版本：邪恶的天才正在与控制卫星的人工智能作斗争，这些人工智能最终被自己的自我压倒。

谁控制您的机器人？ (Who controls your robots?)

Who knows the exact nature of these robots and how they are influencing our lives? Probably no one. When you create an algorithm where machine-learning bootstraps its own knowledge, it is often difficult to determine exactly why some particular decision or prediction was made — it is simply happening at too great a scale for a person to fully understand with limited time.

谁知道这些机器人的确切性质以及它们如何影响我们的生活？ 可能没人。 创建机器学习引导自己的知识的算法时，通常很难确切地确定为什么要做出某些特定的决策或预测-只是发生的规模太大，以至于人们在有限的时间内无法完全理解。

And even without the inherent uncertainly, how can we know whether these language technologies are being used for good rather than evil? These use cases here could as easily be used to have a negative impact on your education, health, relationships, news consumption, and security.

即使没有固有的不确定性，我们如何才能知道这些语言技术是否被用于善良而不是邪恶？ 这些用例很容易被用来对您的教育，健康，人际关系，新闻消费和安全性产生负面影响。

谢谢机器人 (Thank you robots)

While our predictions about robots may have over-emphasized their human form, their ability to speak, and their seemingly autonomous nature, history did get one part right: we need to be aware of how the balance of control is shifting. Each of these examples should be at least a little unsettling, and this awareness is something that we should all be carrying with us.

尽管我们对机器人的预测可能过分强调了人类的形态，说话的能力以及看似自主的性质，但历史确实做到了正确的一部分：我们需要意识到控制的平衡是如何发生变化的。 这些示例中的每一个都应该至少令人有些不安，而这种意识是我们所有人都应随身携带的。

For this awareness, we thank the real creators of these 20 imagined robots: the authors, filmmakers, etc, who have explored the implications of allowing decisions to be made by machines instead of people. We may not have the exact robots that they imagined, but they have made us explore many of implications in advance of the actual technologies. Our part now is to continue to explore the of implications of relying on machines to make decisions for us, especially the ones who increasingly understand what we are saying, even if they don’t talk back.

对于这种意识，我们感谢这20种想象中的机器人的真正创造者：作家，电影制片人等，他们探索了允许由机器而不是由人来做出决定的含义。 我们可能没有他们想象中的确切机器人，但是它们使我们在实际技术之前探索了许多含义。 现在，我们的工作是继续探索依靠机器为我们做出决策的含义，尤其是那些即使不回头也越来越了解我们在说什么的人。

– Rob Munro, 2012

– Rob Munro，2012年

Key to robots (impressed if you identified them all!):

机器人钥匙(如果您全部识别出它们，就会留下深刻的印象！)：

1 HAL 9000, 2001 A Space Odyssey, MGM2 Dalek, Dr Who, BBC3 Robby the Robot, The Forbidden Planet, MGM4 David, Prometheus, 20th Century Fox5 Johnny 5, Short Circuit, TriStar Pictures6 SICO, Rocky IV, MGM/UA7 Motoko Kusanagi, Ghost in the Shell, Manga Entertainment8 Batou, Ghost in the Shell, Manga Entertainment9 Astro Boy, Astro Boy, Mushi Entertainment10 Cyclon, Battlestar Galactica, Universal11 Rosie, The Jetsons, Hanna-Barbera12 Tin Man, The Wizard of Oz, MGM13 Johnny Cab, Total Recall, TriStar14 Google Car, TechCrunch, Google15 Data, Star Trek, Paramount16 The Terminator, Terminator, Orion17 Megatron, Transformers, Hasbro18 Optimus Prime, Transformers, Hasbro19 Huey, Silent Running, Universal20 Bender, Futurama, 20th Century Fox

1 HAL 9000、2001 A Odyssey ，MGM2 Dalek， Who博士 ，BBC3 Robby the Robot， The Forbidden Planet ，MGM4 David， Prometheus ，20世纪Fox5 Johnny 5， 短路 ，TriStar Pictures6 SICO， Rocky IV， MGM / UA7 Motoko Kusanagi ，《 攻壳机动队》 ，Manga Entertainment8巴豆，《 攻壳机动队》 ，Manga Entertainment9，《 Astro Boy》，《 Astro Boy》 ，《 Mushi Entertainment》， 《 10 Cyclon》，《 太空堡垒卡拉狄加》 ，《 Universal11 Rosie》，《 The Jetsons》 ，《 Hanna-Barbera》， 《 12 Tin Man》， 《绿野仙踪》 ，MGM13 Johnny Cab ， Total Recall ，TriStar14 Google Car， TechCrunch ，Google15 Data， Star Trek ，Paramount16 The Terminator， Terminator ，Orion17 Megatron， Transformers ，Hasbro18 Optimus Prime， 变形金刚 ，Hasbro19 Huey， Silent Running ，Universal20 Bender， Futurama ，20世纪福克斯

PS: This article came from a series of talks I gave in 2012 about the dangers of AI. There’s only one public video that I’m aware of, a 5 minute Ignite talk:

PS：本文来自我在2012年就AI的危害进行的一系列演讲。 我只知道一个公共视频，一个5分钟的Ignite演讲：