代理政治：信号和噪点
作者：Hito Steyerl
原标题：Proxy Politics: Signal and Noise
翻译：小饰
A while ago I met an extremely interesting software developer who was working on smartphone camera technology. Photography is traditionally thought to represent what is out there by means of technology, ideally via an indexical link. But is this really true anymore? The developer explained to me that the technology for contemporary phone cameras is quite different from traditional cameras: the lenses are tiny and basically crap, which means that about half of the data being captured by the camera sensor is actually noise. The trick, then, is to write the algorithm to clean the noise, or rather to discern the picture from inside the noise.
不久之前我遇见了一位十分有趣的软件开发商，在智能手机的相机科技领域工作。摄影传统上被认为是再现现实的一种科技手段，且最好是通过一个索引链接。但现在还是这样吗？这个开发商告诉我当代手机的摄像头与传统的相机技术非常不同:手机镜头太小，基本上狗屁不如，这意味着一半被镜头传感器捕获的数据都是噪点。于是关键在于编写代码来消除这些噪点，或者从噪点中识别图像。
But how can the camera know how to do this? Very simple: it scans all other pictures stored on the phone or on your social media networks and sifts through your contacts. It analyzes the pictures you already took, or those that are associated with you, and it tries to match faces and shapes to link them back to you. By comparing what you and your network already photographed, the algorithm guesses what you might have wanted to photograph now. It creates the present picture based on earlier pictures, on your/its memory. This new paradigm is being called computational photography.
但是摄像头是怎么做到的？非常简单：它扫描储存在手机或者社交媒体上所有其他的图片，并在你的联系人中进行筛选。它分析你已经拍过的照片，或者那些与你有关的照片，并试图匹配人脸和形状，将它们与你联系起来。通过比较你和你的网络已经拍过的照片，算法可以推测出你现在可能想拍什么。它根据先前的照片和你/它的记忆，创造出当下的图像。这种新的模式被称为计算性摄影。
The result might be a picture of something that never even existed, but that the algorithm thinks you might like to see. This type of photography is speculative and relational. It is a gamble with probabilities that bets on inertia. It makes seeing unforeseen things more difficult. It will increase the amount of noise just as it will increase the amount of random interpretation.
其结果可能是一张关于根本不存在的东西的照片，但算法认为你可能想看到它。这种摄影是投机性和关系性的。它是一场以惯性为赌注的概率赌博。它使看到未预见的东西变得更加困难。它会持续增加噪点的数量，与此同时也不断地增加随机解释量。
And that’s not even to mention external interference into what your phone is recording. All sorts of systems are able to remotely shut your camera on or off: companies, governments, the military. It could be disabled in certain places—one could for instance block its recording function close to protests or conversely broadcast whatever it sees. Similarly, a device might be programmed to autopixelate, erase, or block secret, copyrighted, or sexual content. It might be fitted with a so-called dick algorithm to screen out NSFW (Not Suitable/Safe For Work) content, automodify pubic hair, stretch or omit bodies, exchange or collage context, or insert location-targeted advertising, pop-up windows, or live feeds. It might report you or someone from your network to the police, PR agencies, or spammers. It might flag your debt, play your games, broadcast your heartbeat. Computational photography has expanded to cover all of this.
更不用提外界对手机记录的干涉。各种各样的系统都能远程关闭你的摄像头：公司、政府、军队。它可以在某些地方被禁用——例如靠近抗议活动的地方被关闭记录功能，或者反过来说监控它看到的一切。同样，一个设备可以被编程为自动像素化、删除或阻止秘密、版权或性内容。它可能装有所谓的鸡巴算法，可以筛选出NSFM（not suitable for work 对工作环境不适宜/不安全的内容），自动地修改阴毛，拉伸或剪裁掉身体，交换或拼贴背景，抑或插入定位广告、弹出窗口或直播。它可能会将你或你的网络中的某人报告给警察、公关机构或垃圾邮件发送者。它可能标记你的债务，玩你的游戏，广播你的心跳。计算性摄影已经扩展到涵盖所有这些。
It links control robotics, object recognition, and machine learning technologies. So if you take a picture on a smartphone, the results are not as premeditated as they are premediated. The picture might show something unexpected, because it might have cross-referenced many different databases: traffic control, medical databases, frenemy photo galleries on Facebook, credit card data, maps, and whatever else it wants.
它将控制机械、物体识别和机器学习技术联系起来。因此，如果你用智能手机拍了一张照片，其结果并不像预设的那样。这张照片可能会显示一些意想不到的东西，因为它可能已经交叉引用了许多不同的数据库：交通管制、医疗数据库、Facebook上的情敌照片库、信用卡数据、地图，以及其他任何它想要的东西。
Relational Photography
关系性摄影
Computational photography is therefore inherently political—not in content but in form. It is not only relational but also truly social, with countless systems and people potentially interfering with pictures before they even emerge as visible.2 And of course this network is not neutral. It has rules and norms hardwired into its platforms, and they represent a mix of juridical, moral, aesthetic, technological, commercial, and bluntly hidden parameters and effects. You could end up airbrushed, wanted, redirected, taxed, deleted, remodeled, or replaced in your own picture. The camera turns into a social projector rather than a recorder. It shows a superposition of what it thinks you might want to look like plus what others think you should buy or be. But technology rarely does things on its own. Technology is programmed with conflicting goals and by many entities, and politics is a matter of defining how to separate its noise from its information.3
因此，计算性摄影在本质上是政治性的—不是在内容上，而是在形式上。它不仅是关系性的，而且是社会性的，无数的系统和人可能会在照片出现之前就对其进行干预。当然，网络并不是中立的，他们代表了司法、道德、审美、技术、商业和被粗暴地隐藏的参数和效果。你可能会在自己的照片中被粉刷，通缉，重定向，征税，删除，被改造或者被替换。相机变成了一个社会的投影仪，而不是一个记录器。它显示的是它认为你可能想要的样子，加上其他人认为你可能认可什么或者成为什么的叠加。但技术很少自己行动。技术是由许多实体以互相冲突的目标编程的，政治决定了如何将其噪点与信息分开。
So what are the policies already in place that define the separation of noise from information, or that even define noise and information as such in the first place? Who or what decides what the camera will “see”? How is it being done? By whom or what? And why is this even important?
所以有哪些决定噪点与信息分离的政策已经实施了？或者说定义这些噪点和信息？谁或者什么决定了相机应该‘看到’什么？它是如何做到的？由谁或者什么决定？那么为什么这一点如此重要？
The Penis Problem
阴茎问题
Let’s have a look at one example: drawing a line between face and butt, or between “acceptable” and “unacceptable” body parts. It is no coincidence that Facebook is called Facebook and not Buttbook, because you can’t have any butts on Facebook. But then how does it weed out the butts? A list leaked by an angry freelancer shows precise instructions given on how to build and maintain Facebook’s face, and it shows us what is well known: that nudity and sexual content are strictly off limits, except art nudity and male nipples, but also how its policies on violence are much more lax, with even decapitations and large amounts of blood acceptable.4
让我们来看一个例子：在脸和屁股之间，或者在“可接受的”和“不可接受的”身体部位之间划一条界线。Facebook被称为Facebook而不是Buttbook并非巧合，因为Facebook上不能有任何屁股。但是，它是如何剔除屁股的呢？一个愤怒的自由职业者泄露的名单显示了关于如何建立和维护facebook网站的精确指示，它向我们展示了众所周知的东西：裸体和性内容是严格禁止的，除了艺术裸体和男性乳头，但它对暴力的政策也更加宽松，甚至斩首和大量的血都是可以接受的。
“Crushed heads, limbs etc are OK as long as no insides are showing,” reads one guideline. “Deep flesh wounds are ok to show; excessive blood is ok to show.”5 Those rules are still policed by humans, or more precisely a global subcontracted workforce from Turkey, the Philippines, Morocco, Mexico, and India, working from home, earning around 4 USD per hour.6 These workers are hired to distinguish between acceptable body parts (face) and unacceptable ones (butts). In principle, there is nothing wrong with having rules for publicly available imagery. Some sort of filtering process has to be implemented on online platforms: no one wants to be spammed with revenge porn or atrocities, regardless of there being markets for such imagery. The question concerns where and how to draw the line, as well as who draws it, and on whose behalf. Who decides on signal vs. noise?
"只要不露出内脏，压碎的头颅、肢体等都是可以的，"一条准则写道。"5 这些规则仍然由人负责，或者更准确地说，由来自土耳其、菲律宾、摩洛哥、墨西哥和印度的全球分包商负责，他们在家工作，每小时收入约4美元。原则上说，对公开的图像制定规则并无不妥。在网络平台上必须实施某种过滤程序：没有人希望收到报复性色情或暴行的垃圾邮件，不管这种图像是否有市场。问题是在哪里以及如何划定界限，以及由谁来划定，代表谁来划定。谁来决定信号与噪点？
Let’s go back to the elimination of sexual content. Is there an algorithm for this, like for face recognition? This question first arose publicly in the so-called Chatroulette conundrum. Chatroulette was a Russian online video service that allowed people to meet on the web. It quickly became famous for its “next” button, for which the term “unlike button” would be much too polite. The site’s audience first exploded to 1.6 million users per month in 2010. But then a so-called “penis problem” emerged, referring to the many people who used the service to meet other people naked.7 The winner of a web contest called in to “solve” the issue ingeniously suggested to run a quick facial recognition or eye tracking scan on the video feeds—if no face was discernible, it would deduce that it must be a dick.8
让我们回到消除性内容的问题上。有没有像人脸识别那样的算法？这个问题第一次公开出现在所谓的Chatroulette难题中。Chatroulette是一个俄罗斯的在线视频服务，允许人们在网上见面。它很快就因其 "下一个 "按钮而闻名，用 "不喜欢按钮 "这个词来形容它就太客气了。2010年，该网站的受众首次爆炸性地达到每月160万用户。但随后出现了一个所谓的 "阴茎问题"，指的是许多人使用该服务与其他人裸体见面。7 一个网络竞赛的获胜者被要求 "解决 "这个问题，他巧妙地建议对视频画面进行快速的面部识别或眼球跟踪扫描--如果没有人脸被识别出来，它将推断这一定是一个阴茎。
This exact workflow was also used by the British Secret Service when they secretly bulk extracted user webcam stills using their spy program, Optical Nerve. Video feeds of 1.8 million Yahoo users were intercepted in order to develop face and iris recognition technologies. But—maybe unsurprisingly—it turned out that around 7 percent of content did not show faces at all. So—as suggested for Chatroulette—they ran face recognition scans on everything and tried to exclude the dicks for not being faces. It didn’t work so well: in a leaked document the GCHQ admits defeat: “there is no perfect ability to censor material which may be offensive.”9
英国特勤局在使用其间谍程序Optical Nerve秘密批量提取用户网络摄像头剧照时也使用了这一确切的工作流程。他们截获了180万雅虎用户的视频资料，以开发人脸和虹膜识别技术。但是--也许并不令人惊讶--事实证明，大约7%的内容根本没有显示人脸。因此，正如对Chatroulette的建议，他们对所有内容进行了脸部识别扫描，并试图排除那些不属于脸部的阴茎。这并不奏效：在一份泄露的文件中，GCHQ承认失败："没有完美的能力来审查可能具有冒犯性的材料 "9。
Subsequent solutions became a bit more sophisticated. Probabilistic porn detection calculates the amount of skin-toned pixels in certain regions of the picture, producing complicated taxonomic formulas.10 But this method got ridiculed pretty quickly because it produced so many false positives, including, as in some examples, wrapped meatballs, tanks, or machine guns. More recent porn-detection applications use self-learning technology based on neural networks, computational verb theory, and cognitive computation. They do not try to statistically guess at the image, but rather try to understand it by identifying objects through their relations.11
随后的解决方案变得更加复杂。概率色情检测计算图片某些区域的肤色像素的数量，产生复杂的分类公式。10但这种方法很快遭到嘲讽，因为它产生了许多假阳性，如在一些例子中，包裹的肉球、坦克或机关枪。最近的色情检测应用程序使用基于神经网络、计算动词理论和认知计算的自学技术。它们并不试图从统计学上猜测图像，而是试图通过识别物体的关系来理解图像。
According to developer Tao Yang’s description, there is a whole new field of cognitive vision studies based on quantifying cognition as such, on making it measurable and computable.12 Even though there are still considerable technological difficulties, this effort represents a whole new level of formalization; a new order of images, a grammar of images, an algorithmic system of sexuality, surveillance, productivity, reputation, and computation that links with the grammatization of social relations by corporations and governments.
根据开发者杨涛的描述，有一个全新的认知视觉研究领域，其基础是将认知量化，使其可测量和可计算。12 尽管仍有相当大的技术困难，这一努力代表了一个全新的形式化水平；一个新的图像秩序，一个图像的语法，一个关于性、监视、生产力、声誉和计算的算法系统，与公司和政府的社会关系语法化相联系。
So how does this work? Yang’s porn-detection system must learn how to recognize objectionable parts by seeing a sizable mass of them in order to infer their relations. So basically you start by installing a lot of photos of the body parts you want eliminated on your computer. The database consists of folders full of body parts ready to enter formal relations. Not only pussy, nipple, asshole, and blowjob, but asshole, asshole/only and asshole/mixed_with_pussy. Based on this library, a whole range of detectors get ready to go to work: the breast detector, pussy detector, pubic hair detector, cunnilingus detector, blowjob detector, asshole detector, hand-touch-pussy detector. They identify fascinating sex-positions such as the Yawning and Octopus techniques, The Stopperage, Chambers Fuck, Fraser MacKenzie, Persuading of the Debtor, Playing of Cello, and Watching the Game (I am honestly terrified of even imagining Fraser MacKenzie).13
那么，这是如何做到的呢？杨的色情检测系统必须学会如何通过看到大量的不良部位来识别它们，以便推断它们的关系。因此，基本上你开始在电脑上安装大量你想消除的身体部位的照片。该数据库由充满准备进入正式关系的身体部位的文件夹组成。不仅有阴部、乳头、屁眼和口交，还有屁眼、屁眼/只有和屁眼/混合_与阴部。在这个库的基础上，一系列的检测器准备开始工作：乳房检测器、阴部检测器、阴毛检测器、舔阴检测器、口交检测器、屁眼检测器、手触阴部检测器。他们识别出迷人的性爱姿势，如打哈欠和章鱼技术、"停止"、钱伯斯操、弗雷泽-麦肯齐、劝说债务人、演奏大提琴和观看游戏（说实话，我甚至害怕想象弗雷泽-麦肯齐）。
*一个身体识别算法试图在这些图像中检测出 "色情"，然后用白框进行审查。*
This grammar as well as the library of partial objects are reminiscent of Roland Barthes’ notion of a “porn grammar,” where he describes the Marquis de Sade’s writings as a system of positions and body parts ready to permutate into every possible combination.14 Yet this marginalized and openly persecuted system could be seen as a reflex of a more general grammar of knowledge deployed during the so-called Enlightenment.
这种语法以及部分对象的图书馆让人想起罗兰-巴特的 "色情语法 "概念，他把萨德侯爵的著作描述为一个准备好的位置和身体部位的系统，可以排列成各种可能的组合。14 然而这种被边缘化和公开迫害的系统可以被看作是所谓启蒙运动期间部署的更普遍的知识语法的反射。
Michel Foucault as well as Theodor W. Adorno and Max Horkheimer compared de Sade’s sexual systems to mainstream systems of classification.15 Both were articulated by counting and sorting, by creating exhaustive, pedantic, and tedious taxonomies. And Mr. Yang’s enthusiasm for formalizing body parts and their relations to one another similarly reflects the huge endeavor of rendering cognition, imaging, and behavior as such increasingly quantifiable and commensurable to a system of exchange value based in data.
米歇尔·福柯以及阿多诺和Max Horkheimer将萨德的性系统与主流的分类系统进行了比较。而杨先生对身体部位及其相互关系的形式化的热情，也同样反映了使认知、成像和行为等日益量化并可与基于数据的交换价值体系相称的巨大努力。
Undesirable body parts thus become elements of a new machine-readable, image-based grammar that might usually operate in parallel to reputational and control networks, but that can also be linked to it at any time. Its structure might be a reflex of contemporary modes of harvesting, aggregating, and financializing data-based “knowledge” churned out by a cacophony of partly social algorithms embedded into technology.
因此，不受欢迎的身体部位成为一种新的机器可读、基于图像的语法的元素，通常可能与声誉和控制网络平行运作，但也可以在任何时候与之相连。它的结构可能是当代模式的反射，即通过嵌入技术的部分社会算法的嘈杂声来收集、聚合和金融化基于数据的 "知识"。
Noise and Information
噪点和信息
But let’s come back to the question in the beginning: What are the social and political algorithms that clear noise from information? The emphasis, again, is on politics, not algorithm. Jacques Rancière has beautifully shown that this division corresponds to a much older social formula: to distinguish between noise and speech to divide a crowd between citizens and rabble.16 If someone didn’t want to take someone else seriously, or to limit their rights and status, one pretends that their speech is just noise, garbled groaning, or crying, and that they themselves must be devoid of reason—and therefore exempt from being subjects, let alone holders of rights. In other words, this politics rests on an act of conscious decoding—separating “noise” from “information,” “speech” from “groan,” or “face” from “butt,” and from there neatly stacks its results into vertical class hierarchies.17 The algorithms now being fed into smartphone camera technology to define the image prior to its emergence are similar to this.
但让我们回到一开始的问题。从信息中清除噪点的社会和政治算法是什么？这里再次强调的是政治，而不是算法。雅克-朗西埃（Jacques Rancière）极好地表明，这种划分对应于一个更古老的社会公式：区分噪音和言论，将人群分为公民和乌合之众。16如果有人不想认真对待别人，或限制他们的权利和地位，人们就假装他们的言论只是噪音、乱七八糟的呻吟或哭泣，而缺乏理性，因此免于成为主体，更不用说权利的拥有者。换句话说，这种政治依赖于一种有意识的解码行为--将 "噪音 "从 "信息 "中分离出来，将 "讲话 "从 "呻吟 "中分离出来，将 "脸 "从 "屁股 "中分离出来，并将其结果整齐地堆积成垂直的阶级等级。17 现在被输入智能手机摄像技术的算法，在图像出现之前对其进行定义，与此相似。
*计算机文件夹将身体部位分类：阴部、乳头、屁眼、屁眼/仅有的、屁眼/混合_与阴部，以及口交，以便学习检测算法。*
In light of Rancière’s proposition, we might still be dealing with a more traditional idea of politics as representation.18 If everyone is aurally (or visually) represented, and no one is discounted as noise, then equality might draw nearer. But the networks have changed so drastically that nearly every parameter of representative politics has shifted. By now, more people than ever are able to upload an almost unlimited number of self-representations. And the level of political participation by way of parliamentary democracy seems to have dwindled in the meantime. While pictures float in numbers, elites are shrinking and centralizing power.
根据朗西埃的主张，我们可能仍然在处理一个更传统的政治即代表的想法。18如果每个人都被听觉（或视觉）代表，没有人被折算为噪音，那么平等可能会越来越近。但是，网络已经发生了巨大的变化，几乎所有代表政治的参数都发生了变化。到现在，比以往任何时候都更多的人能够上传几乎无限数量的自我代表。而通过议会民主方式的政治参与水平似乎在这段时间内有所下降。当图片在数量上浮动时，精英们却在缩减并集中了权力。
And on top of this, your face is getting disconnected—not only from your butt, but also from your voice and body. Your face is now an element—a face/mixed_with_phone, ready to be combined with any other item in the library. Captions are added, or textures, if needs be. Face prints are taken. An image becomes less of a representation than a proxy, a mercenary of appearance, a floating texture-surface-commodity. Persons are montaged, dubbed, assembled, incorporated.
而在这之上，你的脸正在被切断--不仅是与你的屁股，而且与你的声音和身体。你的脸现在是一个元素--脸/混合_与手机，准备与库中的任何其他项目相结合。如果需要的话，可以添加标题，或纹理。脸部指纹被拍摄。一个图像与其说是一个代表，不如说是一个代理，一个外观的雇佣兵，一个漂浮的纹理-表面-商品。人被蒙上了眼睛，被配音，被组合，被纳入。
Humans and things intermingle in ever-newer constellations to become bots or cyborgs.19 As humans feed affect, thought, and sociality into algorithms, algorithms feeds back into what used to be called subjectivity. This shift is what has given way to a post-representational politics adrift within information space.
人类和事物在越来越多的星座中交融，成为机器人或电子人。19 当人类将情感、思想和社会性反馈到算法中时，算法又反馈到过去被称为主体性的东西。这种转变为信息空间中的后代表政治让路。
*Hacer24580318的推特为Flappy Tayyip做广告，这是一款由土耳其总统塔伊普-埃尔多安为主角的游戏。*
Proxy Armies
代理服务军队
Let’s look at one example of post-representational politics: political bot armies on Twitter. Twitter bots are bits of script that impersonate human activity on social media sites. In large synchronized numbers they have become formidable political armies.21 A Twitter chat bot is an algorithm wearing a person’s face, a formula incorporated as animated spam. It is a scripted operation impersonating a human operation.
让我们来看看后代表政治的一个例子：推特上的政治机器人军队。推特机器人是在社交媒体网站上冒充人类活动的一些脚本。在大量同步的情况下，它们已经成为强大的政治军队。21 推特聊天机器人是一种戴着人脸的算法，是一种作为动画垃圾的公式。它是一种冒充人类操作的脚本操作。
Bot armies distort discussions on Twitter hashtags by spamming them with advertisement, tourist pictures, or whatever. They basically add noise. Bot armies have been active in Mexico, Syria, Russia, and Turkey, where most political parties have been said to operate such bot armies. The ruling AKP alone was suspected of controlling 18,000 fake Twitter accounts using photos of Robbie Williams, Megan Fox, and other celebs: “In order to appear authentic, the accounts don’t just tweet out AKP hashtags; they also quote philosophers such as Thomas Hobbes and movies like PS: I Love You.”22
僵尸军团通过用广告、旅游图片或其他东西来扭曲Twitter标签上的讨论。他们基本上是增加噪音。僵尸军队在墨西哥、叙利亚、俄罗斯和土耳其都很活跃，据说这些国家的大多数政党都在运作这样的僵尸军队。仅执政的AKP就被怀疑控制了18,000个使用罗比-威廉姆斯、梅根-福克斯和其他名人照片的虚假Twitter账户。"为了显得真实，这些账户不仅在推特上发布AKP的标签，他们还引用哲学家如托马斯-霍布斯和电影如PS：我爱你。
So who do bot armies represent, if anyone, and how do they do it? Let’s have a look at the AKP bots. Robbie Williams, Meg Fox, and Hakan43020638 are all advertising “Flappy Tayyip,” a cell phone game starring then AK prime minister (now president) Tayyip Recep Erdoğan. The objective is to hijack or spam the hashtag #twitterturkey to protest PM Erdoğan’s banning of Twitter. Simultaneously, Erdoğan’s own Twitter bots set out to detourn the hashtag.
那么，机器人军队代表谁，如果有的话，他们是如何做到的？让我们来看看AKP的机器人。罗比-威廉姆斯、梅格-福克斯和哈坎43020638都在为 "Flappy Tayyip "做广告，这是一款由当时的AK总理（现在的总统）塔伊普-埃尔多安主演的手机游戏。其目的是劫持或发送垃圾标签#twitterturkey，以抗议总理埃尔多安对Twitter的禁止。同时，埃尔多安自己的推特机器人也开始诋毁这个标签。
Let’s look at Hakan43020638 more closely: a bot consisting of a copy-pasted face plus product placement. It takes only a matter of minutes to connect his face to a body by way of a Google image search. On his business Twitter account it turns out he sells his underwear: he works online as an affective web service provider.23 Let’s call this version Murat, to throw yet another alias into the fray. But who is the bot wearing Murat’s face and who is a bot army representing? Why would Hakan43020638 be quoting Thomas Hobbes of all philosophers? And which book? Let’s guess he’s quoting from Hobbes’s most important work, Leviathan. Leviathan is the name of a social contract enforced by an absolute sovereign in order to fend off the dangers presented by a “state of nature” in which humans prey upon one another. With Leviathan there are no more militias and there is no more molecular warfare of everyone against everyone.
让我们更仔细地看看哈坎43020638：一个由复制粘贴的脸和产品定位组成的机器人。通过谷歌图片搜索，只需几分钟就可以将他的脸和身体联系起来。在他的商业推特账户上，我们发现他是卖内衣的：他在网上做情感网络服务提供商。但是，戴着Murat脸的机器人是谁，机器人军队又代表谁？为什么Hakan43020638会引用所有哲学家中的托马斯-霍布斯的话？又是哪本书？让我们猜测一下，他引用的是霍布斯最重要的作品《利维坦》。利维坦是一个社会契约的名字，由一个绝对的君主执行，以抵御人类相互捕食的 "自然状态 "带来的危险。有了利维坦，就不再有民兵，也不再有每个人对每个人的分子战争。
*这是一张机器人的面孔。*
But now we seem to be in a situation where state systems grounded in such social contracts seem to fall apart in many places and nothing is left but a set of policed relational metadata, emoji, and hijacked hashtags. A bot army is a contemporary vox populi, the voice of the people according to social networks. It can be a Facebook militia, your low-cost personalized mob, your digital mercenaries, or some sort of proxy porn. Imagine your photo being used for one of these bots. It is the moment when your picture becomes quite autonomous, active, even militant. Bot armies are celebrity militias, wildly jump-cutting between glamour, sectarianism, porn, corruption, and conservative religious ideologies. Post-representative politics are a war of bot armies against one another, of Hakan against Murat, of face against butt.
但现在我们似乎处于这样一种情况：以这种社会契约为基础的国家体系似乎在许多地方分崩离析，只剩下一套受监控的关系元数据、表情符号和被劫持的标签。机器人军队是当代的vox populi，是社交网络中人民的声音。它可以是Facebook的民兵，你的低成本个性化暴徒，你的数字雇佣兵，或某种代理色情。想象一下你的照片被用于这些机器人中的一个。这是你的照片变得相当自主、活跃、甚至好战的时刻。机器人军队是名人民兵，在魅力、宗派主义、色情、腐败和保守的宗教意识形态之间疯狂地跳动。后代表制政治是一场机器人军队之间的战争，哈坎对穆拉特，脸对屁股。
This may be why the AK pornstar bots desperately quote Hobbes: they are already sick of the war of Robbie Williams (IDF) against Robbie Williams (Electronic Syrian Army) against Robbie Williams (PRI/AAP), they are sick of retweeting spam for autocrats—and are hoping for just any entity organizing day care, gun control, and affordable dentistry, whether it’s called Leviathan or Moby Dick or even Flappy Tayyip. They seem to say: we’d go for just about any social contract you’ve got!24
这可能是AK色情明星机器人拼命引用霍布斯的话的原因：他们已经厌倦了罗比-威廉姆斯（IDF）对罗比-威廉姆斯（叙利亚电子军）对罗比-威廉姆斯（PRI/AAP）的战争，他们厌倦了为专制者转发垃圾信息--希望有任何实体组织日间护理、枪支管制和可负担的牙科，不管它叫利维坦还是莫比-迪克，甚至是Flappy Tayyip。他们似乎在说：我们愿意接受你的任何社会契约！24
Now let us go even one step further. Because a model for this might already be on the horizon. And unsurprisingly, it also involves algorithms.
现在让我们更进一步。因为这种模式可能已经出现在我们的视野中了。而且不出所料，它也涉及到算法。
*这里的代理人是按地区绘制的。*
Blockchain
区块链
Blockchain governance seems to fulfill hopes for a new social contract.25 “Decentralized Autonomous Organizations” would record and store transactions in blockchains akin to the one used to run and validate bitcoin. But those public digital ledgers could equally encode votes or laws. Take for instance bitcongress, which is in the process of developing a decentralized voting and legislation system.26 While this could be a model to restore accountability and circumvent power monopolies, it above all means that social rules hardwired with technology emerge as Leviathan 2.0:
区块链治理似乎实现了对新社会契约的希望。25 "去中心化自治组织 "将在区块链中记录和存储交易，类似于用于运行和验证比特币的区块链。但这些公共数字账本同样可以对投票或法律进行编码。以bitcongress为例，它正在开发一个去中心化的投票和立法系统。26 虽然这可能是一个恢复问责制和规避权力垄断的模式，但它首先意味着以技术为基础的社会规则会作为利维坦2.0出现。
When disassociated from the programmers who design them, trustless blockchains floating above human affairs contain the specter of rule by algorithms ... This is essentially the vision of the internet as techno-leviathan, a deified crypto-sovereign whose rules we can contract to.27
当与设计它们的程序员分离时，漂浮在人类事务之上的无信任区块链包含着算法统治的幽灵......。这基本上是互联网作为技术-利维坦的愿景，一个被神化的加密主权者，而我们可以与他的规则签约。
Even though this is a decentralized process which no single entity at the top controls, it doesn’t necessarily mean no one controls it. Just like smartphone photography, it needs to be told how to work: by a multitude of conflicting interests. More importantly, this would replace bots as proxy “people” with bots as governance. But then again, which bots are we talking about? Who programs them? Are they cyborgs? Do they have faces or butts? And who is drawing the line? Are they cheerleaders of social and informational entropy? Killing machines? Or a new crowd, which we are already part of?28
即使这是一个分散的过程，没有一个实体在顶部控制，但这并不一定意味着没有人控制它。就像智能手机摄影一样，它需要被告知如何工作：由众多相互冲突的利益来决定。更重要的是，这将取代作为代理 "人 "的机器人，以机器人作为治理。但话说回来，我们谈论的是哪些机器人？谁为它们编程？他们是电子人吗？他们有脸还是有屁股？又是谁在划线？它们是社会和信息熵的拉拉队吗？杀人机器？还是一个新的人群，而我们已经是其中的一部分？
Let’s come back to the beginning: How to separate signal from noise? And how does the old political technology of using this distinction to rule change with algorithmic technology? In all examples, the definition of noise rested increasingly on scripted operations, on automating representation and/or decision-making. On the other hand, this process potentially introduces so much feedback that representation becomes a rather unpredictable operation that looks more like the weather than a Xerox machine. Likeliness becomes subject to likelihood—reality is just another factor in an extended calculation of probability. In this situation, proxies become crucial semi-autonomous actors.
让我们回到开头。如何区分信号和噪音？而利用这种区分进行统治的古老政治技术如何随着算法技术的发展而改变？在所有的例子中，噪音的定义越来越依赖于脚本化操作，依赖于代表和/或决策的自动化。另一方面，这个过程可能会引入如此多的反馈，以至于代表权成为一个相当不可预测的操作，看起来更像天气而不是施乐机。相似性变得受制于可能性--现实只是扩展的概率计算中的另一个因素。在这种情况下，代理成为关键的半自主行为者。
*托马斯-霍布斯的《利维坦》（1651年）的封面插图。*
Proxy politics
代理政治
To better understand proxy politics, we could start by drawing up a checklist:
为了更好地理解代理政治，我们可以写一份清单：
Does your camera decide what appears in your photographs? Does it go off when you smile?
你的相机是否决定你的照片中出现什么？当你微笑时，它是否会有提示？ And will it fire in a next step if you don’t?
如果你不笑，它是否会在下一步开火？ Do underpaid outsourced IT workers in BRIC countries manage your pictures of breastfeeds and decapitations on your social media feeds?
金砖四国的低薪外包IT工人是否在你的社交媒体上管理你的母乳喂养和斩首的照片？ Is Elizabeth Taylor tweeting your work?
伊丽莎白-泰勒是否在推特上发表你的作品？ Are some of your other fans bots who decided to classify your work as urinary mature porn?
你的其他一些粉丝是否是决定将你的作品归类为泌尿系成熟色情作品的机器人？ Are some of these bots busily enumerating the names of nation states alongside bodily orifices?
其中一些机器人是否在忙着把民族国家的名字和身体上的口器一起列举出来？
Is your total result something like this?
你的总结果是这样的吗？
(’I`)
(*’σ з｀) ～♪
(’台｀)
(≧∀≦)
(ﾟｪﾟ)(ﾉ∀`)
(/∇＼*)｡o○♡
(/ε＼) (/ε＼)(/ε＼*)
Congratulations! Welcome to the age of proxy politics!
祝贺你! 欢迎来到代理政治的时代!
A proxy is “an agent or substitute authorized to act for another person or a document which authorizes the agent so to act.”29 But a proxy could now also be a device with a bad hair day. A less than authorized agent. A scrap of script caught up in a dress code double bind. A “Persuading the debtor” detector throwing a tantrum over genital pixel probability. A drone gone rogue. Or a delegation of chat bots casually pasting pro-Putin hair lotion ads to your Instagram. It could also be something much more serious, wrecking your life in a similar way—sry life!
代理人是 "被授权代表另一个人行事的代理人或替代人，或授权代理人如此行事的文件。"29 但现在代理人也可能是一个头发乱的装置。一个未经授权的代理人。陷入着装规范双重束缚的废旧剧本。一个 "劝说债务人 "的探测器对生殖器像素的概率大发脾气。一架无人驾驶飞机变得叛逆。或者一个聊天机器人代表团在你的Instagram上随意粘贴支持普京的洗发水广告。它也可能是更严重的事情，以类似的方式破坏你的生活--遗憾的生活！"。
Proxies are devices or scripts tasked with getting rid of noise as well as the bot armies hell-bent on producing it. They are masks, persons, avatars, routers, nodes, templates, or generic placeholders. They share an element of unpredictability—which is all the more paradoxical considering that they arise as result of maxed out probabilities. But proxies are not only bots and avatars, nor are proxy politics restricted to datascapes. Proxy warfare is quite a standard model of warfare—one of the most important examples being the Spanish Civil War. Proxies add echo, subterfuge, distortion, and confusion to geopolitics. Armies posing as militias (or the other way around) reconfigure or explode territories and redistribute sovereignties.30 Companies pose as guerillas and legionnaires as suburban Tupperware clubs. A proxy army is made of guns for hire, with more or less ideological decoration. The border between private security, PMC’s, freelance insurgents, armed stand-ins, state hackers, and people that just got in the way has become blurry. Remember that corporate armies were crucial in establishing colonial empires (East India Company among others) and that the word company itself is derived from the name for a military unit. Proxy warfare is a prime example of a post-Leviathan reality.
代理人是负责摆脱噪音以及一心想要制造噪音的僵尸军团的设备或脚本。它们是面具、人、化身、路由器、节点、模板或通用占位符。它们都有一个不可预测的因素--考虑到它们是由于最大的概率而产生的，这就更加自相矛盾了。但代理不只是机器人和化身，代理政治也不限于数据场景。代理人战争是相当标准的战争模式，其中最重要的例子是西班牙内战。代理人为地缘政治增加了回声、潜规则、扭曲和混乱。军队冒充民兵（或反过来）重新配置或引爆领土，重新分配主权。30 公司冒充游击队，军团冒充郊区特百惠俱乐部。一个代理军队是由雇佣的枪支组成的，或多或少有意识形态的装饰。私人保安、PMC、自由叛乱者、武装替身、国家黑客和只是碍事的人之间的边界已经变得模糊不清。请记住，公司军队在建立殖民帝国的过程中至关重要（东印度公司等），公司这个词本身就来自于一个军事单位的名称。代理人战争是后利维坦时代现实的一个典型例子。
Now that this whole range of activities has long since gone online, it turns out that proxy warfare is partly the continuation of PR by different means.31 Besides marketing tools repurposed for counterinsurgency ops there is a whole range of government hacking (and counterhacking) campaigns that require slightly more advanced skills. But not always. As the leftist Turkish hacker group Redhack reported, the password of the Ankara police servers was 12345.32
现在，这一系列的活动早已上网，事实证明，代理战争在一定程度上是以不同手段进行公关的延续。31除了营销工具被重新用于反叛乱行动外，还有一系列的政府黑客（和反黑客）活动，需要稍微高级的技能。但并非总是如此。正如土耳其左派黑客组织Redhack所报道的，安卡拉警方服务器的密码是1234532。
To state that online proxy politics are reorganizing geopolitics would be similar to stating that burgers tend to reorganize cows. Indeed, just as meatloaf arranges parts of cows with plastic, organic remnants, and elements formerly known as paper, proxy politics position companies, nation states, hacker detachments, FIFA, and the Duchess of Cambridge as equally relevant entities. Those proxies tear up territories by creating netscapes that are partly unlinked from geography and national jurisdiction.
说网上代理政治正在重组地缘政治，就像说汉堡包倾向于重组奶牛一样。事实上，正如肉饼将牛的一部分与塑料、有机残余物和以前称为纸的元素排列在一起一样，代理政治将公司、民族国家、黑客分队、国际足联和剑桥公爵夫人定位为同样相关的实体。这些代理人通过创造与地理和国家管辖权部分无关的网状空间来撕毁领土。
But proxy politics also works the other way. A simple default example of proxy politics is the use of proxy servers to try to bypass local web censorship or communications restrictions. Whenever people use VPNs and other internet proxies to escape online restrictions or conceal their IP address, proxy politics are given a different twist. In countries like Iran and China, VPNs are very much in use.33 In practice though, in many countries, companies close to censor-happy governments also run the VPNs in a exemplary display of efficient inconsistency. In Turkey, people used even more rudimentary methods—changing their DNS settings to tunnel out of Turkish dataspace, virtually tweeting from Hong Kong and Venezuela during Erdoğan’s short-lived Twitter ban.
但是，代理政治也可以从另一个角度发挥作用。代理政治的一个简单的默认例子是使用代理服务器来试图绕过当地的网络审查或通信限制。每当人们使用VPN和其他互联网代理服务器来逃避在线限制或隐藏自己的IP地址时，代理政治就会有不同的变化。在伊朗和中国这样的国家，VPN的使用非常多。33但实际上，在许多国家，与喜欢审查的政府关系密切的公司也在经营VPN，这也是高效不一致的典范。在土耳其，人们使用更初级的方法--改变他们的DNS设置，以隧道方式离开土耳其的数据空间，在埃尔多安短暂的推特禁令期间，几乎从香港和委内瑞拉发推。
In proxy politics the question is literally how to act or represent by using stand-ins (or being used by them)—and also how to use intermediaries to detourn the signals or noise of others. And proxy politics itself can also be turned around and redeployed. Proxy politics stacks surfaces, nodes, terrains, and textures—or disconnects them from one another. It disconnects body parts and switches them on and off to create often astonishing and unforeseen combinations—even faces with butts, so to speak. They can undermine the seemingly mandatory decision between face or butt or even the idea that both have got to belong to the same body. In the space of proxy politics, bodies could be Leviathans, hashtags, juridical persons, nation states, hair transplant devices, moody chat bots, or freelance SWAT teams. Body is added to bodies by proxy and by stand-in. But these combinations also subtract bodies (and their parts) and erase them from the realm of never-ending surface to face enduring invisibility.
在代理政治中，问题就在于如何通过使用替身（或被替身使用）来行动或代表--以及如何使用中介来减弱他人的信号或噪音。而代理政治本身也可以被扭转和重新部署。代理政治将表面、节点、地形和纹理堆叠起来，或者将它们彼此断开。它将身体的各个部分断开，并将它们打开或关闭，以创造出往往是惊人的和不可预见的组合--甚至可以说是有屁股的脸。它们可以破坏脸部或屁股之间看似强制性的决定，甚至是两者都必须属于同一个身体的想法。在代理政治的空间里，身体可以是Leviathans、标签、法人、民族国家、植发设备、情绪化的聊天机器人或自由的特警队。身体通过代理和替身被添加到身体中。但这些组合也减去了身体（和它们的部分），并将它们从永无止境的表面领域中抹去，以面对持久的隐身。
Or maybe something much more simple? In an unprecedented self-experiment I pointed my cell phone at Twitter bot @leyzuzeelizan’s (now deleted) profile picture. With as much authority as I could muster, and hoping it would not shoot me back, I ordered it to run a retina scan on her and send it through its network database. My phone identified her in a split second without a fraction of hesitation. @leyzuzeelizan turns out to be no one other than myself, turned from signal to noise, from face to butt and back again several times over, across the crumbling borders of several nation states and countless levels of towering stacks, erasing differences between bodies, nations, animals, and media containers to advertise the work of someone called Hito Steyerl #oral, #xhamster, #videos, #syria How Not To Be Seen.
或者是更简单的事情？在一次史无前例的自我实验中，我把手机对准了Twitter机器人@leyzuzeelizan（现已删除）的个人照片。我以我所能发出的最大权威，并希望它不会向我开枪，命令它对她进行视网膜扫描，并通过其网络数据库发送。我的手机在一瞬间就识别了她，没有丝毫犹豫。@leyzuzeelizan原来不是别人，而是我自己，从信号变成了噪音，从脸变成了屁股，又从屁股变成了信号，跨越了几个民族国家摇摇欲坠的边界和无数层高耸的堆栈，消除了身体、民族、动物和媒体容器之间的差异，为一个叫Hito Steyerl的人的作品做广告 #oral, #xhamster, #videos, #syria How Not To Be Seen。
In the end, however, a face without a butt cannot sit. It has to take a stand. And a butt without a face needs a stand-in for most kinds of communication. Proxy politics happens between taking a stand and using or being used as a stand-in. It is in the territory of displacement, stacking, subterfuge, and montage that both the worst and the best things happen.[footnote As Tiziana Terranova writes: “A cultural politics of information is crucially concerned with questioning the relationship between the probable, the possible and the real. the cultural politics of information involves a stab at the fabric of possibility, an undoing of the coincidence of the real with the given.” (Network Culture 2004)
然而，最终，一张没有屁股的脸坐不了。它必须采取一种立场。而一个没有脸的屁股需要一个替身来进行大部分的交流。代理政治发生在表明立场和使用或被使用为替身之间。正如Tiziana Terranova所写的："信息的文化政治关键在于质疑可能的、可能的和真实的之间的关系，信息的文化政治涉及到对可能性结构的刺杀，对真实与给定的重合的破坏"。(网络文化 2004)
Also see p 20: “The relation between the real and the probable, however, also evokes the spectre of the improbable, the fluctuation and hence the virtual. As such, a cultural politics of information somehow resists the confinement of social change to a closed set of mutually excluding and predetermined alternatives; and deploys an active engagement with the transformative potential of the virtual (that which is beyond measure).”]
另见第20页。"然而，真实与可能之间的关系也唤起了不可能、波动以及虚拟的幽灵。因此，信息的文化政治在某种程度上抵制了将社会变革限制在一套封闭的、相互排斥的、预先确定的备选方案中；并部署了对虚拟的变革潜力（那是无法衡量的）的积极参与"] 。
注释
1
Rubinstein, Daniel 和 Sluis, Katrina (2013) Notes on the Margins of Metadata; Concerning the Undecidability of the Digital Image. Photographies, 6 (1), 151-158. ISSN 1754-0763（印刷版），1754-0771（在线版）。见→. 另见Katrina Sluis关于这一概念的著作和访谈。
2
关于嵌入到噪音和信息定义中的政治，请参见Tiziana Terranova：《网络文化》，第 "推论Ib：信息的文化政治涉及到回归交流的最低条件（信号与噪音的关系和接触的问题）"。
3
这实际上是引发信息理论的问题，在克劳德-香农（Claude Shannon）1948年发表的一篇开创性论文中。当然，它也是试图设计如何在很多不同的平台上对这些参数进行联网和调制的特点。见Shannon, C.E. (1948), A Mathematical Theory of Communication, " Bell System Technical Journal, 27, 379-423, 623-656, July & October, 1948.→.
4
Adrian Chen, "Facebook外包的反色情和戈尔大队内部，"骆驼趾 "比 "碎头 "更有攻击性。”→.
5
Ibid.
6
他们在家工作，每班4小时，每小时挣1美元，外加佣金（根据招聘信息，这些钱加起来应该是每小时4美元左右的 "目标 "工资）。”
7
Brad Stone, "In Airtime Video Chat Reboot, Nudists Need Not Apply." 布拉德-斯通，"在视频聊天重启中，裸体主义者不适用"。June 5, 2012. →.
8
See →.
9
See →.
10
比如说这个。→ a. 如果皮肤像素相对于图像大小的百分比小于15%，则该图像不是裸体。否则，进入下一步。 b. 如果最大的皮肤区域的皮肤像素数小于总皮肤数的35%，第二大区域的皮肤像素数小于总皮肤数的30%，第三大区域的皮肤像素数小于总皮肤数的30%，则该图像不是裸体。d. 如果总的皮肤数少于图像总像素数的30%，并且边界多边形内的皮肤像素数少于多边形大小的55%，则图像不是裸体。 e. 如果皮肤区域的数量超过60，并且多边形内的平均强度小于0.25，则图像不是裸体。 f. 否则，图像是裸体。
11
美国杨氏科学研究所有限责任公司的视频和图像的色情检测软件（YangSky）→。 →.
12
详见 →.
13
对于有秩序的语法需求的朋友，这里是完整的清单：1. 传教士体位, 侧入式传教士, 2.压扁躺椅, 2.和平手势, 2.蝴蝶体位, 2.阴茎对准技术, 2.止血法, 2.打哈欠体位, 2.章鱼体位, 2.脚踩肩头, 2.狗爬式, 蛙跳式, 直立狗爬式, 展鹰式, 勺子体位, 反向和平手势, 钱伯斯操, 弗雷泽-麦肯锡, 倒立传教士, 2. 女牛仔体位/亚马逊体位, 反向女牛仔/反向亚马逊, 反向女牛仔水平, 亚洲女牛仔→, 2.水平反向, 扶手椅, 黑蜂, 劝说债务人, 拉大提琴, 求婚, 分层, 看球, 反向捎带, 站立和携带, 站立, 手推车等。
14
This is Girish Shambu reading Roland Barthes: Sade Fourier Loyola: “Sades system (according to Barthes), like a language, has its own grammar (“a porno-grammar”), consisting of some basic elements. Sexual posture is the main one, and the others are: sex, male or female; social position; location, e.g. convent, dungeon, even bedroom!, etc. Sade then combines these elements together in all manner of exhaustive permutations to elaborate a fully-fleshed out (sorry) set of possibilities. (Girish Shambu) →.
这是Girish Shambu对罗兰-巴特的解读：《萨德-傅里叶-洛约拉》："萨德系统（根据巴特的说法），像一种语言，有自己的语法（"色情语法"），由一些基本元素组成。性的姿势是主要的，其他的是：性别，男性或女性；社会地位；地点，如修道院、地牢，甚至卧室！等等。然后，萨德将这些元素以各种形式的详尽排列组合在一起，以阐述一套完全充实的（抱歉）可能性。(Girish Shambu) →.
15
启蒙的辩证法
16
雅克-朗西埃（Jacques Rancière）。"关于政治的十篇论文"，载于。 理论与事件。 Vol. 5, No. 3, 2001. (英文)。"为了拒绝将政治主体的头衔赋予一个类别--工人、妇女等......--传统上足以断言他们属于一个'家庭'空间，一个与公共生活分离的空间；在这个空间里，只能出现表达痛苦、饥饿或愤怒的呻吟或呐喊，但不能出现展示共同论点的实际演讲。而这些类别的政治性（......）在于使不可见的东西变得可见；使只能作为噪音听的东西被当作语言来听→.
17
显然，还有其他各种等级制度。
18
Ranciere在1995年的 "La mesentente "中首次阐述了这个想法。从那时起，声音和图像的政治已经随着网络和社会媒体的发展发生了相当大的变化。
19
在唐娜·哈拉维（Donna Haraways）的传奇描述中。半机械人是一种控制论的有机体，是机器和有机体的混合体，是社会现实的生物，也是虚构的生物。唐娜-哈拉维，《赛博格宣言》。科学、技术和二十世纪末的社会主义女权主义"，载于《模拟人、赛博格和妇女》。The Reinvention of Nature (New York; Routledge, 1991), 149-181.
20
Tiziana Terranova区分了表现性空间和信息性空间 网络文化，36
21
使用机器人影响公众舆论的行为被称为 "星际传播"（astroturfing）。""如果社交机器人可以大量创建，它们就有可能被用来偏向公众舆论，例如，通过编写大量的假消息，不诚实地改善或损害公众对某个话题的看法，"论文指出。" 美国国防部在一个名为BotOrNot的可公开访问的在线平台上共同资助了对机器人和非机器人的区分的研究。
22
Elcin Poyrazlar: 土耳其领导人禁止自己的推特机器人军队 发布时间：03/26/14 13:25 EDT。下面的例子是基于Peter Nut和Dieter Leder对土耳其推特机器人军队的研究，在这里引用了其他地方。Elcin Poyrazlar: 土耳其领导人禁止自己的推特机器人军队 发布时间：03/26/14 13:25→.
23
The day is not far when you will be an AK bot too, if you are young and somewhat white, and if you aren’t already.
如果你是年轻的白人，如果你还没有成为AK机器人，那么这一天就不远了。
24
不足为奇的是，西方特务机构似乎也在效仿，对机器人军队进行编程，以便在Fb上自动调整影响。 →.
25
Brett Scott, Vision of a Techno-Leviathan: 比特币区块链的政治。 见 ，6月1日。2014. →.
26
详见 →.
27
Ibid, 25.
28
正如唐娜-哈拉维的《赛博格宣言》已经预言的那样。
29
详见 →.
30
见大卫-里夫关于俄罗斯机器人军队的精彩文字：Cheburashka Fascism？
31
计算机网络话语中 "代理 "和 "代理 "的词源，1998年9月18日。 Joseph Reagle. Revised: January 15, 1999.
32
阿萨德政府的一些服务器似乎也是这种情况。
33