Why Zuboff Was Wrong

Remember how Zuboff insisted that surveillance capitalism wasn’t ideological?

She was completely wrong.

An author I’m reading (full review will be available once I finish the book) conducted a curious experiment.

He entered the words “European art” into Google images. And a strange thing happened. The search returned a large number of paintings depicting black people. European art depicting black people did exist historically. But it was nowhere near half of all European paintings, like the search shows.

Then he entered the words “white couples.” Most of the images showed interracial or black couples. Try it yourself. It’s definitely an experience.

Then he searched for “straight couples.” The author is gay, so he was curious. You can guess what came up. Hint: not only or even mostly straight couples.

He decided to try this experiment in different languages. And a curious thing happened. The answers were the most skewed in English-language searches. They were better in Western European languages. But when he tried the same searches in Turkish, the results were suddenly normal. White couples were white. Straight people were straight. European art was exactly what we think it is. I tried the experiment in Russian, and European art came up as not nearly as… erm, diverse as in the English search. Actually, it wasn’t diverse at all. The Spanish search returned more diversity than the Russian but a lot LOT less than the English search.

It’s very obvious that the search results are manipulated before you can see them. Or maybe there’s a totally benign explanation for why these searches look like a fantasy of the Chief Diversity officer at Reed College. It might be a total coincidence that “the political atmosphere in Silicon Valley is several degrees to the left of a liberal arts college. Social justice activism is assumed – correctly – to be the default setting for all employees in the major companies and most of them, including Google, put applicants through tests to weed out anyone with the wrong ideological inclinations. Those who have gone through these tests recount that there are multiple questions on issues to do with diversity – sexual, racial and cultural – and that answering these questions ‘correctly’ is a prerequisite for getting a job.”

There might be no connection at all and it’s all an innocent glitch.

Or maybe it isn’t.

19 thoughts on “Why Zuboff Was Wrong

  1. The google results have been known forever… I saw the first reference a couple of years ago.

    I just tried it in Polish… “sztuka europejska” gets very… traditional results “biała para” gets a few diverse pairings at the beginning but is mostly…. biała

    The results for “małżeństwo” (marriage or married couple) are very białe as well.

    “Białe małżeństwo” is harder because of the different meanings (it can refer to a marriage that has not been consummated and/or is not intended to be consummated).

    “Straight couple” is harder cause the Polish phrase “para hetero” is awkward and gets a bunch of political stuff rather than actual couples cause ‘hetero’ is only used in political contexts… (public debate on gay issues is handled only in the stupidest manner possible by all sides in Poland).


    1. I’m guessing Poles are a lot less racist and homophobic, according to the defenders’ logic, than English-speakers. And so are Russians and Spanish-speakers. How that’s possible I don’t know but that’s the argument.


      1. \ I’m guessing Poles are a lot less racist and homophobic, according to the defenders’ logic, than English-speakers. And so are Russians and Spanish-speakers. How that’s possible I don’t know but that’s the argument.

        If you are right and the search results are manipulated in this fassion but mainly in English, there are two explanations:

        A- Google workers speak English but don’t know and don’t care about countries like Russia or Poland.

        B – The assumption is that every ‘normal’ person, who is capable of being reformed through re-education, is going to use English at least part of the time. Thus, one doesn’t waste Google’s human resources on hopeless cases who are still stuck in the Middle Ages as evidenced by knowing zero English.

        Besides, Google wouldn’t want to be banned in countries like Saudi Arabia or Iran because of showing one too many gay couples.

        Btw, I searched “гетеросексуальные пары” (heterosexual couples) in Russian and already the third row of results showed a photo of a lesbian couple and following rows showed quite a few gay flags, a lesbian American marriage and a kissing male gay couple that was accidentally showed on Russia_1 while broadcasting Eurovision 2019.


        1. And this is actually a tiny aside in the book I’m reading titled The Madness of Crowds. The author talks in a fascinating way about the inherent conflict between gay people and the queer movement. We can see this split right here on the blog with our gay readers who are fed up with the wokesterism of the queer / trans / fluid crowd. I’m obviously not gay so this is new to me.


          1. There’s always been a rift. Back in the 79s and 80s, many of the most prominent gay activists stood in solidarity with NAMBLA. I was not a gay man during that time, but I’d have to guess that rank and file gay men did not agree with this.


  2. I tried to do the experiment in English and Hebrew. “European art” in Hebrew is

    אמנות אירופאית

    and, unlike the English version which is all paintings, in Hebrew one doesn’t get many paintings unless one clicks on “painting” option above. Instead, there are numerous trifles looking like a huge sale, f.e. a lamp from China or from Europe one can buy in Israel. The very first (and very ugly) search result is (translation from Hebrew) “European and American fashion women creative high heels flowers mural art poster print on canvas home decor picture”

    Even if one clicks on “painting” option the results are not impressive.

    Don’t know whether you can copy paste the Hebrew expression and see yourself.

    If I search “heterosexual couples” or “straight couples”, the results are like in the English version. 🙂 However, “a married couple” or “married couples” both in English and Hebrew present almost only straight couples.

    Btw, how do you find all those wonderful books? Do academics receive special updates? From where? Do you search on Amazon? How do you know whether a book is worth reading? I’ve discovered such good authors on your blog and wonder whether and how one finds them oneself.


    1. I find them myself. I spend a lot of time browsing, looking, searching, both online and in stores.

      You’ve got to read the first 20-30 pages to know if it’s going to be good. Which is why I read a lot more than I finish.


  3. “Chief Diversity officer”

    As you’ve said people choose words for a reason and often can be revealing. Remember this

    “rub the Right’s nose in diversity”

    Those behind this kind of agenda regard the public as animals to be trained and punished when they misbehave. They also despise diversity themselves (or they wouldn’t use a metaphor for animal excrement to describe it)



  4. I do have a benign explanation for “straight couples” and a few other results. May be, Mike or other computer experts on this blog can say whether it is correct and/or explain how the search technology works.

    From what I understand, when I type “straight couples” , Google searches for this expression on different sites \ articles and presents the pictures appearing in those articles. Of course, 99.999% of websites with material about straight couples are not going to use this expression in contrast to web sources talking about gays. That’s why the search engine ‘malfunctions’ in this fashion. It is counterintuitive only to people who have zero idea how computer searches work. In contrast, the expression “married couples” appears in usual, not-LGBT sites all the time and the search results show it. 🙂

    I searched “healthy couples” too and most are straight and quite a few are not (fully) white. Usually the expression appears in articles like “11 Habits of Healthy Couples” and the search showed the accompanying pictures that I expected to see in those kinds of relationship advice articles.

    I think the same thing happens in “straight couples” search regardless of political views of Google employees.


    1. The author absolutely does say that it’s only now allowed to talk about straight couples to show how much worse they are than gay couples.

      Problem is, it’s not massive demand that creates that content. It’s the same tiny social class that does.


  5. A Google worker confirms my words:

    Google Shows You Interracial Couples if You Search for White Couples, Here’s Why

    Apparently when someone uploads a picture of a white couple they don’t use any racial keywords. Basically, white couples are just described as “couples” whereas images of black couples are usually captioned as such. The only time the word “white” tends to get used when describing the couple is when it is an interracial couple, and since Google’s image search looks for keywords it ends up showing you images of interracial couples when you might be looking for a white one. Searching for a couple shows you mostly white couples, which in and of itself shows a racial bias among the users of the internet rather than Google.

    Here are Danny Sullivan’s tweets explaining a bit more about machine learning:


    1. Have you tried entering “married couples”? What’s the first thing you see? Try and you’ll see this guy isn’t telling the truth.

      “You are all racists yourselves” is a tired old kind of rebuttal.


      1. // Have you tried entering “married couples”? What’s the first thing you see? Try and you’ll see this guy isn’t telling the truth.

        The first result gives a couple from the most populated English-speaking country on Earth:

        United States
        Population: 327.2 million (2018)

        Population: 1.339 billion (2017)

        The link from the photo indeed takes one to timesofindia.indiatimes.com

        All couples I noticed were heterosexual and most were white or African / Asian-American, with a few Indian couples thrown into the mix. One couple far below was from Egypt, but there were no Jewish couples.

        Users outside of America also influence search results. Since English is the single international language now, machine learning algorithms are influenced by this world-wide audience in a way the algorithms in Hebrew or other local language are not.

        I am open to your point and the author may be partly right, but one should not ignore the simple explanation of users influencing what a machine learns when the entire system is built on this.

        // The whole business model of Facebook and Google is to get people to pay for other people’s behavior being modified.

        How so? By buying paid services? I have gmail and it is still for free.


      2. —Have you tried entering “married couples”?

        I (in Canada) am seeing various images most of which show couples in actual wedding attire, predominantly white. There is also significant percentage of couples in Indian wedding attire. There are some black couples in the fourth line of images.


  6. The Hebrew example matches a big part of the argument made in Algorithms of Oppression. Noble’s argument there is that Google’s and SEO companies’ ideology is to make money. Weird results come up for some searches because there’s a big enough market for something. Google will fix the problem for specific searches if it becomes a public relations problem for the company, but Google won’t fix the problem throughout the system because fixing it would hurt their reveue stream.


    1. It’s a very silly, unprofessional marketeer who sits and waits for a market to appear organically. The whole point of marketing is to create a market where there is none. The whole business model of Facebook and Google is to get people to pay for other people’s behavior being modified.


  7. Found something interesting and disturbing about gays in Russia (from 2011):

    Сексуальная революция по-русски

    Мы знаем из наших исследований о судьбе российской семьи и отношениях главенства мужа. Тут происходит довольно интересная, хотя и невеселая вещь. В каком-то смысле сохраняется идея, что муж должен приносить заработок, а жена ходить за детьми и так далее. А поскольку экономические и социальные обстоятельства таковы, что в очень многих семьях в России мужья не способны выполнять свою мужскую роль, ну, по экономическим и многим другим причинам, и алкоголизация далеко не на последнем месте, таких мужей или выгоняют, или они становятся в экономическом смысле ничтожными. И возникает новая семья – свободно женская, и из этого вытекают удивительные ручьи, впадающие в два гомосексуальных бассейна. Образуются женские семьи. Это не те люди, которые ходят на парады или готовы сниматься в кино. Я привез эти наблюдения из глубокой российской провинции. Так вот, женщины, уставшие от никчемных мужей, расставшись с ними, далее заключают такие союзы. И это союз в полном смысле семейный, потому что он хозяйственный, он по поводу воспитания детей, создания комфорта. Женщине с женщиной легко, спокойно, хорошо. И возникают сексуальные отношения, но вторичным порядком. И что не менее интересно, наряду с этим возникают мужские союзы, они семьями не являются, но мужчина, который в семье потерпел фиаско, находит другого мужчину, который в таком же положении. Они друг другу не предъявляют претензий, они вместе живут, выпивают, не слушая докучливых баб, которые мешают этому. И возникают устойчивые союзы, пары. Это один край проблематики гомосексуализма.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.