推荐算法竞赛 面向社会_社会操纵算法

推荐算法竞赛 面向社会

Do you know how your apps work? Are you aware of what tech companies are doing in the back with your data? And what’s more revealing: do you know which of your action are actually influenced by those apps? When you take a trip with Uber, buy stuff on Amazon, or watch a movie on Netflix: when are you consciously deciding and when are you being heavily influenced?

您知道您的应用程序如何工作吗？ 您知道数据背后的科技公司在做什么吗？ 而且，更能说明问题的是：您知道哪些操作实际上受这些应用程序影响吗？ 当您乘坐Uber旅行，在Amazon上购买商品或在Netflix上观看电影时：您何时有意识地做出决定，何时受到重大影响？

Tech companies are not passively observing your behavior and acting as a consequence: they are influencing your behavior so you become more predictable. By conditioning your actions, companies can predict outcomes in a better way, and know better what to sell you.

科技公司不是被动地观察您的行为并因此而行事：它们正在影响您的行为，因此您变得更加可预测 。 通过调整您的行动，公司可以更好地预测结果，并更好地知道该卖给您什么。

你的每一次呼吸 (Every breath you take)

Systems and apps make use of massive amount of data like user’s location, maps, browser’s interests, and data streams coming from mobile and wearable devices, in an era of unprecedented power for firms who are no longer merely appealing to our innate desires, but programming our behaviors.

系统和应用程序利用了海量数据，例如用户的位置，地图，浏览器的兴趣以及来自移动和可穿戴设备的数据流，在这个时代，对于那些不再仅仅吸引我们天生的欲望而需要编程的公司来说，力量是空前的我们的行为。

The site True People Search probed the privacy policies of 11 of the biggest tech companies in the world to find out exactly what they know about us, and the result is scary. It’s not the information you are used to share (e.g. your name or email address), but all the information you wouldn’t consider sharing that makes this frightening. Big tech companies record data like income level, political and religious views, credit card information, your calendar events, all your search history and visited websites, and all the content you viewed or engaged with.

True People Search网站调查了全球11家最大的科技公司的隐私政策，以准确了解他们对我们的了解，结果令人震惊。 不是让您共享的信息(例如您的姓名或电子邮件地址)，而是所有您不考虑共享的信息都会使您感到恐惧。 大型科技公司记录的数据包括收入水平，政治和宗教观点，信用卡信息，您的日历活动，所有搜索历史记录和访问过的网站以及您查看或参与的所有内容。

TruePeopleSearch TruePeopleSearch

Uber stores massive amounts of data from its users, including their location, gender, spending history, contacts, phone battery level, whether they are on the way home from a one-night stand, and even if they are drunk or not. Uber has also experimented with its drivers to determine the most effective strategies for keeping them on the road as long as possible.

Uber会存储来自其用户的大量数据 ，包括他们的位置，性别，消费历史记录，联系人，电话电池电量，是否要从一晚看台回家的路上， 即使他们没有喝醉 。 优步还对其驾驶员进行了试验，以确定最有效的策略，以使其尽可能长时间地行驶。

In order to place the right content in front of the right people, Netflix logs everything you have ever watched and how you watch: every time you click, watch, search, play, pause, what programs you consider watching but choose not to, and when you’re most likely to rerun a show. To better identify users’ preferences, content is categorized into tens of thousands of micro-genres, and then paired with a user’s viewing history. Everything you see on Netflix is a recommendation: the rows, the titles in those rows, and the order of those titles within the rows are all deeply considered.

为了将适当的内容展示给正确的人， Netflix会记录您曾经观看的所有内容以及观看方式：每次单击，观看，搜索，播放，暂停，您考虑观看但不选择观看的节目，以及当您最有可能重新播放节目时。 为了更好地识别用户的喜好，将内容分类为数以万计的微型流派 ，然后将其与用户的观看历史记录配对。 您在Netflix上看到的所有内容都是一个建议 ：行，行中的标题以及行中这些标题的顺序都经过深思熟虑。

But Amazon is a data powerhouse taken to a whole new level. They capture absolutely everything, from your product searches, what you look at but don’t buy, what you look at next, how you pay, how you prefer your shipping, your interactions with Alexa, or your requests to Echo. And the shocking thing is the level of detail they store: they capture what device you use, how many items you subsequently clicked on after selecting a product, your location, and the reading sessions and exact time of day for each tap on your Kindle device.

但是亚马逊是一个将数据提升到全新水平的数据强国。 它们绝对捕获了​​所有内容 ，包括您的产品搜索，您所看但不买的东西，您接下来看的东西，付款方式，喜欢运输的方式，与Alexa的互动或对Echo的要求。 而且，令人震惊的是它们存储的详细程度：它们捕获了您使用的设备，选择产品后随后点击了多少项，您的位置以及Kindle设备上每次点击的阅读时间和确切的时间。 。

For Amazon, every mouse click and every twist and turn through its websites, apps and platforms are commodities that carry huge value.

对于亚马逊而言，每一次鼠标单击以及其网站，应用程序和平台的每一次曲折都是具有巨大价值的商品。

When Amazon convinced third-parties to sell their items via their own marketplace, data collection skyrocketed and allowed the company to see way beyond their eye-sight: they can now access to any market they’ve ever wanted to, and see how customers behave in each one of them.

当亚马逊说服第三方通过自己的市场出售商品时，数据收集激增，并使该公司看到了超出其视线的道路：他们现在可以进入他们曾经想要的任何市场，并查看客户的行为方式在他们每个人中。

告诉我我想要什么 (Tell me what I want)

Each day you are influenced by algorithms that guide your decisions and choices. Algorithms are a step by step method for solving a problem or reaching a goal, based on taking an input and conducting a sequence of specified actions to reach a result. Since the explosion of modern technologies, they have expanded, sophisticated and replicated everywhere, having a central role in places like social media platforms.

每天，您都会受到指导决策和选择的算法的影响。 算法是一种逐步解决问题或达到目标的方法，其基础是接受输入并执行一系列指定操作以达到结果。 自从现代技术爆炸以来，它们已扩展，完善并在各处复制，在社交媒体平台等地方起着核心作用。

The goal of several social media and content selection algorithms is to maximize clicks. They are designed to show or recommend stuff that will increase the probability of users clicking on it, since clicking is what generates revenue for the platforms.

几种社交媒体和内容选择算法的目标是最大化点击量 。 它们的设计目的是显示或推荐可以增加用户单击它的可能性的内容，因为单击可以为平台带来收入。

For example, a click-through optimization algorithm is more profitable if it can better predict what people are going to click on, so it can feed them exactly that. So, a way to optimize the result is to feed users with content they like, and don’t show anything outside their comfort zone. Although it’s true that this causes their interests to become narrower, it’s not that the algorithms are trying to show you the stuff you like: they’re trying to turn you into a predictable clicker, taking you to a “predictable point” and making it easier for companies to perform any action (e.g. sell you something).

例如，点击型优化算法可以更好地预测人们将要点击的内容，因此可以为他们提供准确的信息，从而带来更大的利润。 因此，一种优化结果的方法是向用户提供他们喜欢的内容，而不在其舒适区域之外显示任何内容。 尽管这确实使他们的兴趣缩小了 ，但这并不是算法试图向您展示您喜欢的东西：他们试图将您变成可预测的点击器，将您带到“可预测的点”并使其公司更容易执行任何行动(例如，向您出售东西)。

Companies have figured out that they can do this using your own data, by gradually modifying or emphasizing your preferences with targeted material. That’s basically, if you think of a spectrum of preferences, it’s to one side or the other because they want to drive you to an extreme, where you become a more predictable clicker and so they can monetize you more effectively. This is advanced applied behavioral science, or as Jeff Hammerbacher (the founder of Cloudera) said:

公司已经发现，可以通过使用目标材料逐步修改或强调您的偏好来使用您自己的数据来做到这一点。 基本上， 如果您想到了一系列偏好设置，那么它偏向另一侧，因为它们希望将您推向极致 ，使您成为更可预测的点击器，从而使他们可以更有效地从中获利。 这是先进的应用行为科学，或者正如Jeff Hammerbacher ( Cloudera的创始人)所说：

“The best minds of my generation are thinking about how to make people click ads. That sucks.”

“我们这一代最好的人正在考虑如何使人们点击广告。 糟透了。”

The reasons behind this are mainly economic. A Wall Street Journal investigation found that Google manipulated search algorithms to prioritize large businesses over smaller ones, guiding search users to more prominent businesses over lesser-known ones. According to the investigation, eBay saw its traffic from Google nosedive during 2014, contributing to a $200 million decrease in its revenue guidance for the year. After Google told the company it had made a change that lowered the ranking of several high traffic pages, eBay considered pulling its roughly $30 million quarterly advertising spend from Google, but ultimately decided to start a pressure campaign on executives. Google eventually agreed to boost the ranking of several eBay pages that were demoted while eBay took on a costly revision of its web pages to make them more relevant.

其背后的原因主要是经济方面的。 《 华尔街日报》(Wall Street Journal)的一项调查发现，谷歌操纵搜索算法以将大型企业优先于小型企业，从而将搜索用户引导至知名度更高的企业而不是鲜为人知的企业。 根据调查， eBay在2014年从Google获得的流量激增，导致该年度收入指南减少了2亿美元 。 谷歌告诉该公司做出了改变，降低了一些高流量页面的排名后，eBay考虑从谷歌撤出其大约3000万美元的季度广告支出，但最终决定对高管发起压力运动。 Google最终同意提高被降级的几个eBay页面的排名，而eBay对其网站进行了昂贵的修订以使其更具相关性。

Mechanized intervention is an ideal way to keep content flowing towards more lucrative topics, avoiding material that doesn’t generate engagement or profits. For tech companies to succeed, algorithms must focus on monetizable activities, and this is exactly what they are doing with our data.

机械化干预是保持内容流向利润更高主题的一种理想方法，避免了不会产生参与度或利润的材料。 为了使科技公司成功，算法必须专注于可盈利的活动，而这正是他们对我们的数据所做的。

Interested in these topics? Follow me on Linkedin or Twitter

对这些主题感兴趣？ 在Linkedin或Twitter上关注我

翻译自: https://medium.com/swlh/algorithms-of-social-manipulation-1915b5cd339a

推荐算法竞赛 面向社会