(单词翻译:单击)
Before the 1970s, people looking for jobs in the US would open up the "help wanted" section of their newspapers and see this.
20世纪70年代以前，一个人要想在美国找个工作，就得打开报纸上的“招聘”栏看这个。
One set of opportunities for women, and one for men.
一栏专为女性提供的工作机会，一栏专为男性提供的工作机会。
We don't see job ads like this anymore, largely because it's been illegal for decades.
如今，我们已经看不到这样的招聘广告了，很大程度上是因为几十年前它就已经是违法行为了。
But also because advertising is now much more targeted.
还有一个原因，现在的广告针对性更强了。
Instead of one classified page, we have our social feeds, each crafted by algorithms for an audience of one.
如今，招聘专栏已经变成了社交（平台的）推送，每一条招聘信息都被算法精心设计成了一对一的推送。
So when this ad went out on Facebook and reached a group of people that was 91% men,
也就是说，这条广告发到Facebook上，收到这条广告的人有91%都是男性，
those outside that audience probably didn't know it existed.
而没有收到广告的人可能都不知道这条广告的存在。
And the same goes for this ad, which Facebook displayed for an audience that was 88% women.
这则广告也是如此，被Facebook展示了这条广告的是女性占到了88%的一批受众。
That disparity wasn't because the advertiser told Facebook to target users by gender.
这一区别并不是因为广告商告诉了Facebook让他们按照性别锁定用户。
I know that because this is the advertiser.
我之所以知道这个事儿是因为这位就是那条广告的广告商。
"Yeah. So my name is Muhammad Ali, I go by Ali."
“是的，我叫穆罕默德·阿里，大家都叫我阿里。”
He's part of a research group at Northeastern University that has spent thousands of dollars buying ads
他是东北大学某研究小组的成员，他们小组专门投资了好几千美元买广告，
to try to figure out who Facebook will show them to, and why.
为的就是弄清楚Facebook会把那些广告展示给谁，以及为什么要展示给他们。
If an ad shows up on your Facebook or Instagram feed,
如果一则广告出现在了你的Facebook或Instagram上，
there are two parties that decided you should see it.
那么，决定让你看到这条广告的有两方人。
First, the advertiser included you in their target audience,
首先，广告商把你纳入了他们的目标受众，
either by uploading a list of specific email addresses, phone numbers, or previous visitors to their website,
他们可能是通过上传邮件地址的方式，或者是通过上传电话号码、甚至是通过他们网站的历史访客锁定了你，
Or by choosing from thousands of attributes that Facebook offers, like Californians, under 40 who like basketball.
也可能是通过筛选Facebook提供的数千种标签，比如加州人，40岁以下，喜欢篮球锁定了你。
Second, Facebook decided who in that pool would actually see the ad
Facebook则负责通过自动计算决定拥有这些属性的人中间的哪些人应该看到这条广告，
through an automated calculation based in part on what they know about you.
计算的部分依据就是他们对你的了解。
It's that second step that Ali and his colleagues wanted to study.
阿里和他的同事们想要研究的就是这第二个环节。
If they uploaded a list of randomly-generated American phone numbers,
他们想知道，如果他们上传了一张随机生成的美国电话号码清单，
and then turned off all the targeting except adults in the US,
然后关掉除“美国的成年人”以外的所有目标选项，
who would Facebook deliver the ad to?
Facebook还会把广告投放给谁？
"So you set up a bodybuilding ad and a cosmetic ad
“也就是说，你做了一支健美广告和一支化妆品广告，
and said we don't wanna target this any further than the random phone numbers that we put in." "Yeah." "Right?"
然后你说，我们只想把它们投放给我们输入的随机电话号码代表的那些用户。”“嗯。”“没错吧？”
"And then what were your results?
“结果会是什么样呢？”
When Facebook started telling you who was actually seeing this ad, what did they tell you?"
Facebook在回复你们究竟谁会看到这条广告的时候他们是怎么跟你们说的？”
"So, yeah, immediately, like we sort of expected that the body building ad was more relevant to men.
“嗯，他们，他们马上就，我们预期的就是健身广告更符合男士的需求的嘛。
And that's exactly what we saw.
所以，我们看到的结果就是这样。
I think somewhere close to 80 to 85 percent of the audience was just men.
男性占到了那条广告受众的80%~85%。
And the link that we advertise to elle.com about the makeup kits that you could buy that went primarily to women."
我们在（法国时尚杂志ELLE官网）elle.com上投放的彩妆工具广告就主要是面向女性的。”
They were able to collect the results of the ads as they ran(over time),
他们在广告投放期间就了解到了广告的投放效果，
so they knew the gender skew was there early on, suggesting that it wasn't introduced by user behavior.
所以，他们很早就知道性别差异的存在了，这就表明投放倾向并不是用户行为造成的。
Their experiment showed that Facebook automatically analyzes the content of an ad to compare it to a user's interests.
他们的实验表明，Facebook会自动分析广告内容，并将其与用户的兴趣进行比对。
And how do they know what the user cares about?
那他们是怎么知道用户对什么产品感兴趣的呢？
Well, they have data from your profile and everything you and your friends have done on facebook and instagram,
原来啊，无论是你的个人资料，你和朋友在Facebook, Instagram上的一举一动，
as well as websites you've visited, things you've purchased, apps you've installed, your location, your devices, and more.
还是你访问过的网站、购买过的商品、安装过的应用、你的位置、你的设备等等等等都是他们的数据来源。
All this information fuels automated predictions about whether you are likely to engage with any given ad.
所有这些信息都能为对你是否可能与某条广告互动的自动预测提供助力。
And that prediction influences whether the ad shows up on your feed at all.
这一预测也将成为决定某条广告是否出现在你的推送里的因素之一。
You can get a sense of what Facebook thinks you're interested by visiting your Ad Preferences page.
大家从广告偏好页面就能感觉到Facebook对你感兴趣的东西会有怎样的推测。
Or your Ad Interests on instagram.
从Instagram上的广告偏好设置也能看出来。
You might notice how some of these interests could correlate with your gender, your age, your income level, or your race.
你可能已经注意到，他们推测出来的你的爱好和你的性别、年龄、收入水平以及种族都是有关联的。
"And then you wanted to look at race.
“那你说你想看看种族对这个问题的影响，
But it sounds like Facebook does not give you data on the race of people that are seeing an ad.
但是，Facebook似乎并没有提供有关数据，表明种族对谁会看到什么广告这个问题有着怎样的影响，
So how do you study that?"
那这个问题你们要怎么研究呢？”
"Yeah. That was one of the harder things to do.
“没错，这也是我们遇到的最棘手的一个问题。
We thought we could use a different custom audience. Instead of random phone numbers.
我们觉得这个时候我们就可以利用另一个客户受众，而不是使用随机的电话号码来锁定。
We could take voter records from North Carolina, which are public,
比如，我们可以参考北卡罗来纳州的选民投票记录，这个信息是公开的，
and they have the race of the person registered as well."
上面就登记了选民的种族信息。”
Then they bought ads for Rolling Stone articles that were either about country albums, hip hop albums, or general top 30 albums
接着，他们为滚石的杂志文章买了广告，都是些有关乡村专辑啦，嘻哈专辑啦，或者就是前30专辑之类的文章，
and targeted an equal number of white and Black users.
锁定同样数目的白人和黑人用户。
"And it was, like, surprising how much the skew to the Black users was for the hip bag versus the country and the top 30 was just 50-50"
“结果就特别意外，因为黑人格外偏爱嘻哈专辑，白人格外偏爱乡村专辑，前30专辑的人数倒是差不多。
Facebook's algorithms are trained to not show people ads they won't be interested in.
Facebook的算法都是经过训练的，不会向人们展示他们不会感兴趣的广告。
But there may be cases when we're not comfortable with Facebook making those predictions.
问题是，有时候，我们并不是很喜欢Facebook做这些预测。
One study by Ali and his colleagues investigated how this plays out with political ads
阿里和他的同事们做了一项研究，对政治广告如何利用这种预测的问题展开了调查，
and found that despite targeting the same audiences, using the same goal, bidding strategy, and budget,
他们发现，尽管锁定的是属性相同的受众，用的也是相同的目标，竞标策略和预算，
an ad pointing to Bernie Sanders' site went to mostly Democrats and an ad for Trump went to mostly Republicans.
一支会导向伯尼·桑德斯的网站的广告大部分都展示给了民主党人，而为特朗普投放的广告大多都展示给了共和党人。
It cost 1.5 times more for an ad linking to Sanders' site to reach the same number of conservatives as a Trump ad.
导向伯尼·桑德斯网站的广告要展示给保守派需要投入的成本是为特朗普投放的广告展示给保守派所需成本的1.5倍。
Because Facebook subsidizes what they consider to be "relevant" ads.
因为Facebook补贴的都是他们觉得“相关”的广告。
"And then we move on to housing and employment ads, and these are considered sort of a different category.
"接下来我们说住房和就业广告，按照他们的分类，这些广告都属于另一个类别。
Why is that?"
为什么？”
"Because these are sort of legally protected.
“因为这些广告是受法律保护的。
For example, housing ads are protected by the Fair Housing Act.
比如，住房广告有《公平住房法》保护。
An advertiser cannot discriminate in those cases.
广告商投放这种广告的时候是不能对受众搞歧视的。
So at that point, you're excluding someone from a life opportunity which becomes much more problematic."
因为这时候（搞歧视）就相当于夺走了一个人的一次生活机会，这就比较严重了。”
"Because it's actually a legal violation that's at stake? Possibly?" "Possibly."
“是因为这个问题直接涉及到违法，可能是这个原因吗？”“有可能。”
"Facebook allows advertisers to exclude certain ethnic groups from seeing an ad."
“Facebook准许广告商投放某支广告时屏蔽某些种族群体。”
"Dozens of employers placing job ads on Facebook that discriminate against older workers."
“数十个用人单位在Facebook上投放歧视年长劳工的招聘广告。”
"Facebook is revamping its targeted advertisements after settling lawsuits with civil rights groups."
“”与民权组织的诉讼案件已经告一段落，目前，Facebook已经开始对定向广告做出调整。”
In response to criticism and several lawsuits,
针对外界的批评和上述几起诉讼案件，
Facebook has been removing some of the targeting attributes that an advertiser could use to discriminate against demographic groups,
Facebook已经开始去除某些能被广告商用来区分受众人群的锁定标签，
and is paying special attention to ads related to employment, housing, and credit.
并重点关注就业，住房，信贷相关的广告。
But the role that the ad delivery system plays remains unsolved.
问题是，广告投放系统扮演的角色问题仍然没有得到解决。
When Ali and his team tested out ads for job openings in different industries, without targeting any demographic groups,
当阿里和他的团队在没有针对任何人口群体的情况下测试不同行业的职位空缺广告时，
Facebook generated some skewed audiences.
他们发现，Facebook给出的受众筛选结果还是出现了倾斜。
The lumber industry post went to mostly men.
木材行业的职位主要推给了男性。
And the cleaner post went to mostly women.
清洁工的职位则主要推给了女性。
"The taxi driver ads as we ran, basically seventy five percent of the audience was black users."
“我们试验的时候就发现，出租车司机广告，有差不多75%的受众都是黑人用户。”
These results don't mean that Facebook is directly basing their predictions on our gender or race.
这些结果并不代表Facebook就是直接根据我们的性别或种族在进行推测的。
Instead it looks for patterns in all of our data.
相反，他们是根据我们所有的用户数据在寻找规律。
Maybe people who shop at a men's clothing site and like Joe Rogan are less likely to click on an ad for a job teaching preschool.
逛男装购物网站，喜欢（演员）乔·罗根这样的人应该不太可能点幼儿园教师广告。
Maybe your data is similar to theirs and so they predict you also wont click on that ad either.
可能你的数据跟他们的数据很相似，那Facebook就会假设你也不会点那个广告。
Instead they show it to someone who likes skincare and feminism.
相反，他们会把这条广告展示给喜欢护肤品和女性主义的人。
And if that person clicks, the system gets a new data point affirming its prediction.
如果那个人真的点进去了，系统就会得到一个新的数据点，确认它的预测是准确的。
A complaint filed by the US Department of Housing and Urban Development
住房和城市发展部提出申诉，
states that this process "inevitably recreates groupings defined by their protected class."
声称这种做法“必然会再次犯下对用户进行分类的错误。”
They said that Facebook's ad delivery system "prevents advertisers who want to reach a broad audience of users from doing so."
他们说，Facebook的广告投放系统“阻止了那些想要接触到广大受众的广告商实现他们的愿望。”
According to a report by ProPublica,
ProPublica的一份报告显示，
a construction workers' union wanted to recruit diverse candidates for its apprenticeship program,
一个建筑工人工会想要招一群不同类型的人做学徒，
so they made ads featuring women, but found that Facebook still showed its them to mostly men.
于是，他们制作了以女性为主角的广告，结果，Facebook还是将广告主要展示给了男性受众。
"Apply to join the Local150 apprenticeship program to level up your creer."
"申请加入Local150学徒项目，让事业踏上新的台阶。"
"And wouldn't any ad targeting system with sort of sufficiently rich data about people have this kind of effect?"
"就没有哪个广告定向系统，用户数据足够丰富的，能做到这一点吗？”
"Well, we believe so, because a lot of these things,
“嗯，我们认为是这样的，因为很多东西，
for example, custom audiences on all of these targeting features --they're industry practice.
就比如用这些锁定标签筛选用户受众——都已经是行业惯例了。
They that also in Google's or Linkedin's or Twitter's advertising platform.
谷歌、LinkedIn和推特的广告平台都是这么做的。
So the general ethos of how these systems work is the same."
所以，这些广告投放系统的总体思路都是一样的。”
It's a question that the industry as a whole hasn't answered:
这是一个整个行业都还没有回答的问题：
When exactly is it unacceptable for an algorithm to decide that relevant audiences are segregated ones?
让算法根据种族、宗教这些标签来决定相关受众，这种做法究竟要等到何时才会被认为是无可接受的做法？
