One of the more chilling technologies of our present internet age is the deepfake —audio and video imitations of actual people, via face and sound manipulation, that appear real. The most realistic CGI so far has had the “uncanny valley” effect, alarming the senses with a too-close-to-reality-without-getting-there appearance. But when, not if, this tech advances enough to reliably trick the human ears and eyes, we could be due for far more sinister outcomes.

我们现在的互联网时代的更让人心寒的技术ØNE是实际的人的deepfake -audio和视频模仿，通过面部和声音操纵，出现真正的。 迄今为止， 最现实的 CGI具有“不可思议的山谷”效果，以太接近真实而没有出现的外观警告了感官。 但是，如果不是(如果不是)这项技术的发展足以可靠地欺骗人的耳朵和眼睛的话，那么我们就有可能获得更多可怕的结果。

The problem was most vividly illustrated in Jordan Peele’s fabricated PSA from a deepfake Barack Obama about fake news. If you look closely you can see it’s a fake — still, we should expect the technology is only going to improve. And since so many are predisposed to believe more obvious fakes flooding the internet already, what new levels of slander and misinformation will they believe and act on thanks to a more slick presentation as the tech comes along? When the proliferation of videos that show people doing and saying things those people didn’t do or say, the backlash will be inevitable.

这个问题在乔丹·皮尔(Jordan Peele) 伪造的巴拉克·奥巴马(Barack Obama) 关于假新闻的伪造PSA中得到了最生动的说明。 如果您仔细观察，您会发现它是伪造的-仍然，我们应该期望技术只会得到改进。 既然有如此多的人倾向于相信已经有更多明显的假货泛滥到互联网上，那么随着技术的发展，人们将相信并采取什么新的诽谤和错误信息行动，这要归功于更加精巧的演示方式。 当显示人们在做和说他们没有做或说过的事情的视频激增时，这种反弹是不可避免的。

Perhaps the solutions will hinge on counter technologies and systems of public trust: verification processes matching new footage to reputable publishing outlets, digital badges of credibility from ethically-conscious media, detecting and flagging known fakes.

可能的解决方案将取决于对等技术和公众信任系统：将新素材与信誉良好的发行渠道相匹配的验证过程，具有道德意识的媒体的信誉数字徽章，检测和举报已知的假货。

How optimistic you are about these defenses working as a bulwark against deepfakes probably depends on whether you think their current parallels are effective.

您对这些防御措施能否抵御深层造假的乐观程度可能取决于您是否认为当前的类似措施是否有效。

Number me with Sonia Zawitkowski and those who find fact checking and other reality-based verification methods inadequate. Zawitkowski writes:

请给我索尼亚·扎维特科夫斯基(Sonia Zawitkowski)和那些发现事实检查和其他基于现实的验证方法不充分的人。 扎维特科夫斯基写道：

There are now over a dozen websites dedicated to telling people which information they should believe. Take for example one of the most recent ones specific to coronavirus — Infotagion.com, with headlines at the time of writing such as “Can a Televangelist Cure COVID-19?” and “Did COVID-19 Originate on a Comet?”

现在有十几个网站致力于告诉人们应该相信哪些信息。 例如，针对冠状病毒的最新病毒之一-Infotagion.com，在撰写本文时已成为标题，例如“电视传播家可以治愈COVID-19吗？”。 和“ COVID-19起源于彗星吗？”

Picture someone who sincerely believes their parishioner has been imbued by God with healing powers, or a person that skipped right over pangolins and the Wuhan lab theories to the extra-terrestrial. Now imagine them visiting this website and immediately succumbing to its authority. “I was a firm believer in Pastor Grifty’s ability to apostolically denounce the coronavirus from people’s bodies but now, after encountering this fact check, I realize that I was holding onto a mistaken hypothesis!” It’s almost too silly for words.

想像一下一个人，他们真诚地相信自己的教区居民已经被上帝赋予了治愈的能力，或者一个人跳过穿山甲和武汉实验室的理论而直接进入了地球。 现在想象他们访问该网站并立即屈服于其权威。 “我坚信Grifty牧师有能力使人体内从冠状病毒中脱出冠状病毒，但现在，在遇到这一事实检查后，我意识到我坚持了一个错误的假设！” 言语简直太傻了。

It would be nice if a stamp of falsehood would send any given piece of multimedia into the wastebin of discourse, but we know better. Reality isn’t the driving force behind our view of the world, and especially not of the opinions we formulate about our political opponents. We are adept at disbelieving the real.

如果虚假的标记会将任何给定的多媒体内容发送到话语的垃圾箱中，那将是很好的，但是我们知道的更多。 现实不是我们对世界的看法的推动力，尤其不是我们对政治对手提出的观点的推动力。 我们擅长于否认真实。

The Rayshard Brooks killing caught on tape is real, but there’s still plenty of polarizing views about what to make of it. It features the kind of raw intimacy of a horrific moment that a deepfake mad scientist would hoist triumphantly as lightning cracked in the background. The scene is captured at multiple of angles and distances, and every second of the 40-plus minute encounter is viewable. We have a clear accounting of the initial violation, the arrest attempt, and the killing itself. All accepted as real.

录音带上的雷莎德·布鲁克斯杀人案是真实的，但对于如何制作它，仍然有很多分歧的看法。 它的特点是那种恐怖的瞬间的原始亲密感，深深的疯狂科学家会在雷电劈开的背景下胜利地升起。 可以从多个角度和距离捕获场景，并且可以看到40多分钟的相遇中的每一秒。 我们对最初的违规行为，逮捕企图和谋杀本身有清楚的说明。 所有被接受为真实的。

So is there a clear accounting of what the video means? Hardly.

那么，有关视频的含义是否清楚？ 几乎不。

A quick search through coverage of the footage (let alone the interpretations from take artists posted on social media) returns polarized explications firmly condemning or exonerating the actions of the police officers’ actions.

快速搜索 通过该镜头覆盖范围(更不用说从张贴在社交媒体拿艺人的解释)返回极化吃茶坚决谴责或者免除警察的行动的行动。

These waves of interpretation always crash in, whether it’s video of Brooks, Ahmaud Arbery, John Crawford III, or, more recently, George Floyd.

无论是布鲁克斯(Brooks)，艾莫德(Ahmaud Arbery)，约翰·克劳福德(John Crawford)III还是最近的乔治·弗洛伊德(George Floyd)的录像带，这些解释浪潮总会崩溃。

Whatever our valid fears for deepfakes and misinformation, they won’t outpace the recording of actual events anytime soon. The invention of the camera phone has introduced exponentially more real footage of events themselves, heaping corroboration and contextualization of an isolated moment like never before. The second, third, and fourth angles that cops have reconstructed at television crime scenes using a patchwork of evidence and conflicting eyewitness accounts (a thumbprint here, a velocity times force calculation there) are now readily viewable from a preponderance of cameras — on the chests of police officers, from the vantage of a business’s roof corner, and in the hands of scattered bystanders instinctively pressing record.

无论我们对假冒和错误信息的真正担心如何，它们都不会很快超过实际事件的记录。 照相手机的发明引入了事件本身的指数级更多真实镜头，堆积了孤立时刻的确凿和上下文，这是前所未有的。 现在，可以通过大量摄像机在胸部轻松地看到警察在电视犯罪现场使用的证据拼凑而成的第二个，第三个和第四个角度，并且目击者的说法相互矛盾(此处是指纹，此处是速度乘以力的计算)。来自企业屋顶角落的警力，以及分散的旁观者本能地紧迫的记录。

Juries of centuries past would leap at the chance to see so much video footage as they deliberated a case. Introducing such evidence into those settings would vaporize or corroborate many defenses and accusations swiftly. But the would-be juror, attorney, and judge scrolling their feed today is prepared for a heightened amount of information, and conditioned to absorb it into their priors.

几个世纪以来，陪审团在审议案件时会抓住机会观看这么多的录像。 将这些证据引入这些环境将Swift蒸发或证实许多辩护和指责。 但是，今天将要担任陪审员，律师和法官的他们的提要滚动显示是为获取更多信息而做的准备，并且有条件将其吸收到他们的先验知识中。

Not so long ago, a cautious response to incomplete but apparently damning video may have been “let’s see what the other facts are.” Now, the skeptic urges, “Let’s see the rest of the video before making up our mind.” The wheels are already turning. Everyone is his own juror, every news digester her own media member. Add the benefit of the doubt to this frame, remove it from the next.

不久前，对不完整但显然令人讨厌的视频的谨慎React可能是“让我们看看其他事实是什么”。 现在，持怀疑态度的人敦促：“让我们下定决心先看视频的其余部分。” 车轮已经在转动。 每个人都是自己的陪审员，每个新闻摘要者都是她自己的媒体成员。 将疑问的好处添加到此框架中，然后从下一个框架中删除。

That should provide clarity, but often just adds to the layers of opinion and interpretation that aligns with the viewer’s preconceptions. There’s never enough to convince the holdouts. We won’t, in fact, believe it when we see it.

这应该提供清晰度，但通常只会增加与观看者的先入为主的观点和解释层。 说服坚持派永远是不够的。 实际上，当我们看到它时，我们不会相信它。

Not only is the Brooks video itself, frame by frame, dissected in service to a conclusion, there remains a hunger for more footage, or more ornamental information to bolster, what really happened on screen. There is skepticism over which angles are released at specific junctures of the encounter by the police for public framing. What would he have done had he evaded, continuing his desperate path away from the officers, and out of view of the cameras? Slumped over on the sidewalk after the adrenaline rush wore off, easily apprehended? Slumped over the wheel of a stolen vehicle crashed into another driver in the midst of his panicked flee? What expositional detail — Brooks’s rap sheet and assumed level of threat, or the officers’ training and understanding of use of force policy — should we read into the heads of the acting officers?

布鲁克斯视频本身，一帧一帧地被剖析以得出结论，还渴望获得更多的镜头或更多的装饰性信息来支撑屏幕上真正发生的事情。 人们怀疑警察在相遇的特定时刻释放角度以进行公共取景。 如果他逃避了他的绝望之路，远离军官，又离开了摄像机，他会怎么做？ 肾上腺素激流消逝后，容易跌落在人行道上吗？ 陷入被盗车辆的车轮，在他惊慌失措的逃亡中撞向另一位驾驶员？ 我们应该读一下演说人员的头脑中的哪些论述细节—布鲁克斯的说唱表和所假设的威胁程度，或者军官对使用武力政策的训练和理解？

Much has been written on the conspiratorial mind. Bizarre reaches for alternative explanations of the real (like President Trump’s reckless theory of an elderly Antifa “provocateur”) are and will remain confounding, especially as technology makes fake footage more and more plausible. But the way people subtly refuse to agree on the substance of a video like Brooks’s reveals a deeper problem — we don’t need help in deceiving ourselves.

关于阴谋论的思想已经写了很多。 怪异的人提出了关于真实的替代解释(例如特朗普总统对鲁Anti的安提法“ proocateur”的鲁ck 理论 )现在并将继续令人困惑，尤其是随着技术使假镜头变得越来越合理时。 但是，人们巧妙地拒绝就布鲁克斯这样的视频的实质达成共识的方式揭示了一个更深层次的问题-我们在欺骗自己时不需要帮助。

Contrary to hyperbolic criticism of protesters seeking “mob justice,” it is encouraging that their actual demands for what they view as modern lynchings are explicitly a jury trial for perpetrators of brutality. But the outcomes of those trials, and the schemes of state attorneys pursuing charges, remain in question. A verdict of “not guilty,” or “guilty” of a lesser charge, can’t satisfy the passion for justice.

与对寻求“暴民正义”的示威者的夸张批评相反，令人鼓舞的是，他们对他们认为是现代私刑的实际要求明确是针对残暴行径的陪审团审判。 但是，这些审判的结果以及寻求起诉的州律师的计划仍然存在疑问。 对“无罪”或较低罪名的“有罪”判决不能满足对正义的热情。

The methods we’ve extolled in American history and tradition — a jury by trial of peers, informed by forensics experts, human testimony, and legal operators protecting the rights of all — are sufficient in theory. But we can only accept them with an underlying social contract. If we don’t share the hope for transcendent values to be grasped at, the means to those ends are beside the point.

从理论上讲，我们在美国历史和传统中所采用的方法(由同行审判陪审团，法医专家，人类作证和保护所有人权利的合法经营者)是足够的。 但是，我们只能通过基本的社会契约来接受它们。 如果我们不希望抱有超越价值的希望，那么达到这些目的的手段就不重要了。

Our peers in the jury box are not first introduced to the highest stakes of evidence and veracity in a courtroom, but every single day on their social media feeds. We are self-trained to look with our political third eye instead of as impartial observers. Weeding out those biases has always been a necessity in jury selection.

陪审团中的同僚不会首先在法庭上被介绍到最重要的证据和真实性，而是每天在社交媒体上获取信息。 我们是自我训练的，以第三只政治眼看待自己，而不是公正的观察者。 消除这些偏见一直是陪审团选择的必要条件。

It usually just takes bias to get some of us to reject video evidence, not high-tech fakes. In the mind of the technology watcher, deepfakes pose a special threat to our capacity to trust sources of evidence. But our problems won’t start with deepfakes — our problems are already here. Even when we’ve had ubiquitous footage we’ve found ourselves incapable of forming a common understanding of what’s happened. The era of deepfakes isn’t going to thrust us into an epistemological quagmire for the simple reason that we are already in one.

通常，我们中的某些人会拒绝提供视频证据，而不是高科技的假货，这需要偏见。 在技​​术观察者看来，深造假对我们信任证据来源的能力构成了特殊的威胁。 但是，我们的问题不会从假货开始-我们的问题已经在这里。 即使我们拥有无处不在的镜头，我们仍然无法对所发生的事情形成共识。 虚假的时代不会因为我们已经处在一个简单的原因而将我们逼入认识论的泥潭。