The Manipulators Page 5
Facebook Betting on AI
Zuckerberg and other executives have admitted that Facebook is counting on artificial intelligence (AI) to enforce its speech policing.56 Silicon Valley insiders I spoke to were deeply skeptical that AI could better police “hate speech” than people. If the company can’t figure out what’s hate speech and what isn’t using human censors, it will be considerably more difficult for the company to teach AI to accomplish the (oftentimes nuanced) task, those insiders told me. The end result, they said, will be algorithms erring on the side of censorship, and Facebook will have an easy rebuttal against charges of political bias: it’s not us doing the censoring, it’s just the algorithm.
Facebook’s Supreme Court of Speech
Similarly, Facebook hopes to deflect any blame for censorship by creating a sort of “Supreme Court” to oversee its speech policing on the site. As Zuckerberg told left-wing website Vox in April 2018:
[W]hat I’d really like to get to is an independent appeal. So maybe folks at Facebook make the first decision based on the community standards that are outlined, and then people can get a second opinion. You can imagine some sort of structure, almost like a Supreme Court, that is made up of independent folks who don’t work for Facebook, who ultimately make the final judgment call on what should be acceptable speech in a community that reflects the social norms and values of people all around the world.57
The beauty of this system for Facebook is that it could both choose the members of this Supreme Court and deny any responsibility for their decisions.
Shifting responsibility is important for Facebook, because it would help it avoid such embarrassing incidents as its censorship of the Liberty County Vindicator. The Vindicator has served as the local newspaper for the residents of Liberty, Texas, (population 9,215), since 1887. Leading up to July 4, 2018, the Vindicator divided the Declaration of Independence into a series of twelve Facebook posts, posting one per day. On July 2, the tenth day of the Vindicator’s series, Facebook removed the post. As the Texan paper dryly informed its readers: “Somewhere in paragraphs twenty-seven to thirty-one of the Declaration of Independence, Thomas Jefferson wrote something that Facebook finds offensive.”58 Only after the Vindicator published an article on the censored founding document did Facebook restore its post, which it once again chalked up to an inadvertent mistake. The offending paragraphs contained some of American revolutionaries’ grievances against the King of England:
He has abdicated Government here, by declaring us out of his Protection and waging War against us.
He has plundered our seas, ravaged our Coasts, burnt our towns, and destroyed the lives of our people.
He is at this time transporting large Armies of foreign Mercenaries to compleat the works of death, desolation and tyranny, already begun with circumstances of Cruelty & perfidy scarcely paralleled in the most barbarous ages, and totally unworthy the Head of a civilized nation.
He has constrained our fellow Citizens taken Captive on the high Seas to bear Arms against their Country, to become the executioners of their friends and Brethren, or to fall themselves by their Hands.
He has excited domestic insurrections amongst us, and has endeavoured to bring on the inhabitants of our frontiers, the merciless Indian Savages, whose known rule of warfare, is an undistinguished destruction of all ages, sexes and conditions.
“While the Vindicator cannot be certain exactly what triggered Facebook’s filtering program, the editor suspects it was most likely the phrase ‘Indian Savages,’ ” the paper’s managing editor, Casey Stinnett, informed readers. “Perhaps had Thomas Jefferson written it as ‘Native Americans at a challenging stage of cultural development’ that would have been better.” Stinnett and the Vindicator handled the incident well, but that they had to handle anything at all only underscores the disturbing extent to which Facebook has already embraced its role as a global censor and enforcer of political correctness. If posting America’s founding documents to Facebook requires going through an appeal process, we’re already squarely in Orwellian territory.
Facebook’s changes to its rules about employees’ conduct provide an illuminating glimpse into the future. The company issued a memo in January 2019 banning employees from trying to change each others’ minds about politics or religion.59 “We’re keeping it simple with three main guidelines: Don’t insult, bully, or antagonize others. Don’t try to change someone’s politics or religion. Don’t break our rules about harassing speech and expression,” Facebook chief technology officer Mike Schroepfer wrote in the memo.60 “These guidelines apply to all work communications including Workplace, email, chat, tasks, posters, whiteboards, chalkboards, and face-to-face,” Schroepfer added. “Since Workplace is where most of these discussions happen, we are investing engineering resources there.” If Facebook can’t figure out how to have civil discussion within its own workplace without cries for censorship, it can’t possibly make it work on a platform with 1.25 billion daily active users.61 Knowledgeable Facebook sources predict that its speech policies for employees will become its speech policies for users, and controversial opinions—as defined by left-wing activists—will simply be off-limits.
CHAPTER FOUR One Nation under Google
Google has amassed more power, on a global scale, than any corporation in the history of the world. The tech behemoth and its subsidiary, YouTube, provide the results for more than ninety percent of searches for information on the Internet.1 “Google” has become a verb, synonymous with “research.” If you’re looking for an article, you Google it. If you want information about a political candidate, you Google it. If you want to know more about Google itself, you Google it. And to locate a video, you turn to Google-owned YouTube. Google’s control of the flow of information—including information about politics and culture—throughout the world is unprecedented. Even more disturbing is the fact that Google is openly moving towards censorship while remaining opaque about how that censorship occurs.
The news itself flows through Google. Media companies depend on Google—both Google News and Google Search—for a sizable portion of their online traffic. The data analytics firm Parse.ly maintains a monthly list of the top “referrers” of online articles. Every month, one referrer stands head and shoulders above the rest: Google’s search engine. Google Search drove 30.1 percent of article referrals in December 2018, according to Parse.ly. That was more than double its closest competitor, Facebook, which drove 14.5 percent of article referrals.2 The third highest: Google News, with 2.3 percent. No other referrer had more than two percent of the market.
“I want you to imagine walking into a room, a control room with a bunch of people, a hundred people, hunched over a desk with little dials, and that that control room will shape the thoughts and feelings of a billion people. This might sound like science fiction, but this actually exists right now, today,” former Google design ethicist Tristan Harris said in an April 2017 TED Talk.3 “I know because I used to be in one of those control rooms. I was a design ethicist at Google, where I studied how do you ethically steer people’s thoughts? Because what we don’t talk about is how the handful of people working at a handful of technology companies through their choices will steer what a billion people are thinking today,” Harris said. When Google is seeking to “ethically” steer thoughts, it matters a heck of a lot what Google employees think is ethical—and what thoughts they want you to think.
Robert Epstein, a senior research psychologist at the American Institute for Behavioral Research and Technology, has studied how Google’s search results and recommendations shape users’ political opinions. In a study of 661 participants from forty-nine states and ranging in age from eighteen to thirty-five, Epstein found that Google’s “SSE [‘Search Suggestion Effect’] appears to have the power to change a 50/50 split in preferences among people who are undecided on an issue to a 90/10 split.”4 According to the study, “it is possible that the outcomes of upwards of 25 percent of the world’s national elections are
now being determined by Google’s search algorithm, even without deliberate manipulation on the part of company employees.” Based on this data, Epstein and his colleagues concluded that Google’s search algorithm “almost certainly ends up favoring one candidate over another in most political races, and that shifts opinions and votes.” Epstein’s research was prominently featured in the documentary film The Creepy Line, which explored Google’s ability to manipulate thoughts—and elections.
Google doesn’t even need to go so far as downranking specific results about candidates for it to steer users’ views towards one candidate and away from another. Thanks to the dependable left-wing biases of corporate media companies like CNN and the Washington Post, Google only needs to downrank certain outlets that don’t conform to the left-wing narrative, and the outcome will be much the same. The political left know this—it’s why they’ve pressured Google to take those very measures.
Google’s employees are well aware of their own power, as evidenced by leaked internal discussions on how best to wield that power against conservatives. Google’s workplace culture is dominated, top to bottom, by an obsession with left-wing identity politics, lavishing praise on internal teams with a majority of women, and shaming and denigrating internal teams with a majority of men.5 Google incentivizes employees’ attendance at what it calls “anti-bias sessions” where presenters lecture employees on white male privilege.6 (A 2017 study by Altheas Nagai at the Center for Equal Opportunity found that such sessions are ineffective at best,7 and there is some data suggesting that they may in fact be counterproductive.8) PowerPoint slides from a “Bias Busting” session listed as off-limits: “Debating whether bias exists at your organization.”9 Employees are expected to attend the re-education sessions in order to demonstrate their commitment to diversity and inclusion—though of course it’s neither real diversity nor real inclusion. At Google, diversity doesn’t apply to diversity of thought, and inclusion doesn’t apply to orthodox Christians or social conservatives. Diversity at Google means that managers explicitly try to fill quotas that label employees by race and sex. “Diverse” job candidates—non-white people (except for Asians)—are placed in special queues for priority hiring.10 “I could care less about being ‘unfair’ to white men. You already have all the advantages in the world,” one Google employee wrote on an internal message board.11 Opposing different standards for different races, another employee wrote, is a “microaggression.”12 Another employee claimed that Republicans had declared “war” on her by voting for Trump.13
Lower and mid-level employees pressure executives to do more for left-wing causes, both personally and in their professional capacities. I obtained internal Google documents showing that after Trump’s election, some Google employees used the company’s internal message boards to call for a company-wide boycott of all Trump hotels and Trump-linked businesses. “I think that if Google’s company philosophy is to promote inclusion, then to continue to support Trump business endeavors, which are explicitly non-inclusive, would be very contrary to this mission,” one employee explained. It was a popular idea. Google’s left-wing activists have only gotten bolder with time. More than 1,100 Google employees signed an open letter in November 2019 demanding that the company voluntarily adopt essentially its own version of Ocasio-Cortez’s Green New Deal.14 Under the demanded terms, Google would have to somehow reduce its carbon emissions to zero by the year 2030. The company would also have to boycott any “contracts to enable or accelerate the extraction of fossil fuels.”
“Google’s culture has many similatiries to the culture of academia, so it’s hardly surprising that the outrage mobs (and other trends on college campuses) have caught on much faster at Google,” former Google engineer Mike Wacker told me. “It used to be that you would only hear buzzwords and abstract theories such as privilege, critical theory, and intersectionality in the left-wing corners of academia. Now, these ideas have become prevalent at Google. They are often taught as if they are the truth, not just one possible theory, and they have even found their way into documents about performance management,” Wacker added.
Some aspects of the company’s internal culture could be mistaken for parodies of left-wing campus culture. It’s pervaded with snooty elitism and childish temper tantrums. One Google employee from West Virginia recounted conversations denigrating West Virginians and advocating against poor people having children.15 Sources provided me with internal documents showing Google employees having full-fledged meltdowns about the use of the word “family”—which was presumed to be offensive to people without children. One employee stormed out of a March 2017 meeting when a speaker “continued to show (awesome) Unicorn product features which continually use the word ‘family’ as a synonym for ‘household with children.’ ” The employee posted an extended rant, which was well-received by his colleagues, on why linking families to children is “offensive, inappropriate, homophobic, and wrong.”
This is a diminishing and disrespectful way to speak. If you mean “children,” say “children”; we have a perfectly good word for it. “Family friendly” used as a synonym for “kid friendly” means, to me, “you and yours don’t count as a family unless you have children.” And while kids may often be less aware of it, there are kids without families too, you know.
The use of “family” as a synonym for “with children” has a long-standing association with deeply homophobic organizations. This does not mean we should not use the word “family” to refer to families, but it means we must doggedly insist that family does not imply children.
Even the sense, “suitable for the whole family,” which you might think is unobjectionable, is totally wrong too. It only works if we have advance shared conception of what “the whole family” is, and that is almost always used to mean a household with two adults, of opposite sex, in a romantic/sexual relationship, with two or more of their own children. If you mean that as a synonym for “suitable for all people” stop and notice the extraordinary unlikelihood of such a thought! So “suitable for the whole family” doesn’t mean “all people,” it means “all people in families,” which either means that all those other people aren’t in families, or something even worse. Use the word “family” to mean a loving assemblage of people who may or may not live together and may or may not include people of any particular age. STOP using it to mean “children.” It’s offensive, inappropriate, homophobic, and wrong.
Roughly one hundred other Google employees “upvoted” that post, signaling their approval for the rant. Some employees echoed their displeasure with the term. “Thanks for writing this. So much yes,” one employee agreed. “Using the word ‘family’ in this sense bothers me too,” wrote another employee, who felt excluded by the term because she was neither married nor a parent. “It smacks of the ‘family values’ agenda by the right wing, which is absolutely homophobic by its very definition,” she wrote, adding that “it’s important that we fix our charged language when we become aware of how exclusionary it actually is. As a straight person in a relationship, I find the term ‘family’ offensive because it excludes me and my boyfriend, having no children of our own.” Another wrote: “My family consists of me and several other trans feminine folks, some of whom I’m dating. We’re all supportive of each other and eventually aspire to live together. Just because we aren’t a heterosexual couple with 2.5 kids, a white picket fence, and a dog doesn’t mean we’re not a family.”
Yet another employee wrote that “using ‘family’ to mean ‘people with kids’ is also annoying to me as a straight-cis-woman who doesn’t have or want kids. My husband, my parents, and my pets are my family.” Another: “As someone [aromantic-asexual] (and thus both perpetually, intentionally single and unlikely to ever have a child) I appreciate this being pointed out. My family is not incomplete because it has no children and will never have romantic (or more) relationships. It’s incredibly frustrating to constantly run into assumptions that being single and childless is an undesirable, in
complete state—even if those assumptions are implicit and from various statements about what families look like.” Google vice president Pavni Diwanji then joined the conversation and acknowledged that the use of the term “family” had sparked “concerns.” “Hi everyone, I realize what we said at tgif might have caused concerns in the way we talked about families. There are families without kids too, and also we needed to be more conscientious about the fact that there is a diverse makeup of parents and families,” Diwanji wrote. “Please help us get to a better state. Teach us how to talk about it in an inclusive way, if you feel like we are not doing it well. As a team we have a very inclusive culture, and want to do right in this area. I am adding my team here so we can have open conversation,” Dwiwanji concluded.