Careless People: The Book Meta Doesn’t Want You to Read

Careless People by Sarah Wynn-Williams could almost be one of Christopher Buckley’s Beltway satires. Like Thank You for Smoking or The White House Mess, the first-person protagonist takes the reader on a journey from dream job to absurd nightmare—each chapter an ironic critique of the powerful characters depicted. Except Wynn-Williams is real, and so are the truly awful people and events she describes. “…like watching a bunch of fourteen-year-olds who’ve been given superpowers and an ungodly amount of money, as they jet around the world to figure out what power has bought and brought them,” she writes in the prologue.

The subtitle, A Cautionary Tale of Power, Greed, and Lost Idealism aptly describes this memoir, which begins with Wynn-Williams’s story of surviving a shark attack at the age of 13 in her native New Zealand and ends with her being escorted by security from the shark-infested headquarters at Facebook twenty-five years later. Hired in 2011 as the Manager of Global Public Policy, Wynn-Williams conveys her initial enthusiasm as a true believer in the power of Facebook to be a force for good and, on that basis, how she pitched the idea of a policy role for herself at a time when the leadership did not yet grasp why the company would need to build relationships with state leaders.

Initially, Wynn-Williams’s expertise as a former New Zealand diplomat reads like a satirical counterpoint to the fumbles of tech-nerds who don’t understand state craft. An early chapter, for instance, describes the visit of German delegates to Facebook’s Washington office and their bewilderment upon seeing the open-plan office with all the facades stripped away to expose the ducts and bare fixtures to “symbolize” the company’s nascent status. “‘You dismantled the furnishings of a proper office to make it look like this? Like it is under construction?’ one of the officials inquired, incredulous,” Wynn-Williams writes.

This image of the deadpan German thinking he is meeting with unserious people would be funny if not for the very real and deadly events that are indeed foreshadowed. As the narrative unfolds like a thriller, the protagonist discovers unbounded arrogance, callousness, hypocrisy—and ultimately—dangerous and criminal conduct among her superiors. The faux feminism of Sheryl Sandberg and lechery of Joel Kaplan become subplots about elite executives whose worst crime against humanity, so far, is arguably Facebook’s role in fostering rampant hate-speech which fueled the Rohingya genocide in Myanmar between October 2016 and January 2017.

As discussed in an earlier post, Senators Hawley et al., motivated in part by Wynn-Williams’s testimony and accounts in the book, have stated an intent to investigate Facebook’s misconduct designed to appease the Chinese Communist Party. But to me, the most compelling part of the memoir is the glimpse into Mark Zuckerberg’s character, especially as a putative oligarch in context to the Trump-led assault on the constitutional order of the United States.

Wynn-Williams’s portrait of Zuckerberg, an avatar of Big Tech leaders, combines the patriarchal vanity of John Galt with the innocent savagery of Jack Meridew—a boy billionaire, who plays board games that his staff let him win, but who ultimately embraces the destructive power he controls. Specifically, the chapters describing Zuckerberg’s psychological process upon learning that Facebook was catalytic to the 2016 election of Donald Trump can be described as denial, anger, pride, and corruption.

During a flight on the private jet to Lima for the Asia-Pacific Economic Cooperation (APEC) summit, Elliot Schrage, VP of global communications, marketing, and public policy, explains to Zuckerberg how, “A Trump operative named Brad Parscale ran the operation together with embedded Facebook staff, and he basically invented a new way for a political campaign to shitpost its way to the White House, targeting voters with misinformation, inflammatory posts, and fundraising messages,” Wynn-Williams writes.

Initially, Zuckerberg clings to the belief that his platform is a neutral conduit for free speech and “connecting people,” but he then becomes angry at the irrefutable evidence presented by Shrage. Then, at the APEC summit, Zuckerberg’s incipient sense of his own power, and test of his character, is described by Wynn-Williams as he is buffeted between foreign leaders kissing his ass one minute and President Obama in a side meeting lecturing him about the dangers of misinformation on Facebook.

Rather than introspection, Zuckerberg responds like a petulant comic book villain—so offended by the criticism of the U.S. President that he decides to use the power of his technology for his own run at the office. “After all, not only does Mark now have Trump’s playbook, he owns the tools and sets the rules,” Wynn-Williams writes. “And he has something no one else has, the ability to control the algorithm with zero transparency or oversight.”

Again, the image of the staff reacting to Zuckerberg’s announcement that he wants to hold events in swing states like Iowa, New Hampshire, Pennsylvania et al. would make great satire but for the fact that, as Wynn-Williams puts it, “He could run for president and not ask anyone for a dime.”

Of course, the real point is not the prospect of President Zuckerberg—at least not yet—but rather Wynn-Williams’s courageous exposure of the mindset behind the allegedly “greatest tool for democracy ever invented.” And she does so at tremendous personal risk–threatened by Meta, which tried to stop publication of the book, tried to stop her testifying before Congress this month, and threatens to sue her for $50,000 per negative comment about the company.

In many ways, Careless People reveals what many of us already knew about Meta and the other social media giants—at least since 2017: that they are not designed or operated according to principles that ever justified the populist rhetoric of “democratization.” That was a lie more than a decade ago, and the lie is exponential in the battle over development and application of artificial intelligence. Wynn-Williams sums it up well with her thoughts about the travesty in Myanmar:

“I’ve spent a lot of time thinking about what unfolded next in Myanmar, and Facebook’s complicity. It wasn’t because of some grander vision or any malevolence toward Muslims in the country. Nor lack of money. My conclusion: It was just that Joel, Elliot, Sheryl, and Mark didn’t give a fuck.”

eXodus: Bluer Skies for Social Media or Just a Short Breath of Fresh Air?

As of today, the social media platform BlueSky has grown to about 25 million users, which is still a fraction of the 600 million on X, but the recent spike at the former is attributable to people abandoning the latter. After Elon Musk acquired and rebranded Twitter, fired the accountability team, reinstated Trump, and then devoted both X and personal resources to supporting that campaign, the election was the final straw for many who fled to bluer skies.

Built as a “decentralized” platform, BlueSky takes an approach often advocated by Mike Masnick (who sits on the board) as a way to rescue the good of social media from the bad. But as I have argued for more than a decade, much of the harm caused by social media is too subtle to be designed out of the system. Even the best (or best-intended) social platforms are simply bad for democracy. BlueSky’s decentralized architecture may be more effective at weeding out haters and disinformation campaigns and providing users with greater control over what they see, block, etc., but this changes nothing about the reasons social platforms are fundamentally hazardous.

I am just about 50 pages into David Golumbia’s posthumously published magnum opus, Cyberliberatarianism: The Right-Wing Politics of Digital Technology, and one view David and I share is that social media’s organic harms to democratic institutions simply outweigh its benefits as a “social” forum. We discussed this in the podcast we recorded in October 2021 and generally agreed that there is no technological solution for many of the medium’s inherent pitfalls. My short list of those pitfalls includes the following:

Provocative Nonsense Isn’t Just a Joke

It is natural to post and share short-attention-span editorial material like memes. Some of the best educated people I know post this kind of content all the time, and once in a while, I like or share the ones I find funny or on point. But when the subjects of these micro-editorials are political and provocative, they are not wholly distinguishable from “Q drops” as fuel added to a fire. Perhaps the most dismaying example of this is the profusion of memes applauding, or at least winking at, Luigi Mangione for allegedly shooting United Healthcare CEO Brian Thompson.

A subject worthy of its own post (including a report that Mangione made a ghost gun ), the point worth summarizing in this post is that social media has eroded the moral barriers to political violence. Political violence is an inevitable biproduct of the erosion of democratic norms because a party’s destructive conduct (and indeed UH has blood on its hands) invites violence about which many will feel at least ambivalent, if not enthusiastic. Of course, as Dr. King warned, violence only multiplies violence, but as a novel feature of this vicious cycle, social media offers dopamine hits of provocative nonsense that allows even the observers of violence to laugh at, feel self-righteous about, or at least excuse conduct that should be rejected as a principle of functional democracy.

Aggregated Narcissism is the Ignorance of Crowds

Known formally as the Dunning-Kruger effect, microdoses of “information,” combined with the enticement to comment and share, feed that human frailty which allows us to pretend to know more about a particular subject than we do. Then, solidifying and amplifying our ignorance, social media provides “connection” to others who share the same uninformed belief. Thus, while many of us look aghast at the kind of unqualified nutjobs the next administration would tap for leadership of important departments, we must also acknowledge it is not Trump fans alone who have abandoned the notion of expertise concurrent with the growth of digital technology and social media.

Ignorance on topics ranging from vaccinations to NATO is just as deeply rooted in “progressive” politics as the right wing, and social media feeds the beast, partly due to the “IKEA Effect.” Akin to Dunning-Kruger, the IKEA Effect describes the satisfaction derived from completing a DIY project, only instead of assembling a desk without cracking any veneer, social media promotes and rewards the project of doing one’s own research, even to arrive at a conclusion that may be wholly untethered to reality. These psychological effects cannot be “programmed out” of the medium or countered with fact-checking. At best, they can be understood, much as my generation learned to understand the effects of watching too much television.

A Community of Frenemies is not a Community

“Jealousy” and “faction” are two words that appear with great frequency in the founders’ writings advocating adoption of the Constitution and creation of the United States. Whether the subject is election procedure, national defense, taxation, etc., The Federalist and other seminal writings all warn against faction as inherently destructive to common purpose, and out of that debate evolved the tradition of compromise and collaboration as necessary for keeping the Republic. But today, infighting among likely political allies is rampant thanks to social media, and it would be a mistake to believe that the mechanisms at work in the hostile takeover of the GOP are unique to the right-wing.

Although Golumbia presents an excellent case that Silicon Valley ideologies have always been grounded in right-wing, even fascistic, principles—and that bros like Zuckerberg and Musk have intentionally tilted the game in that direction—even “organic” interactions reveal that a prominent individual on the left will be attacked in a hate-storm if she critiques some unfounded position held by “progressives.” Thus, regardless of where people claim to sit on the political spectrum, one result of social media has been to scorn the idea of collaboration itself—a folly which has now become self-fulfilling prophecy because the reelection of an anti-democratic administration justifies the anti-collaborative spirit from which it drew power in the first place.

Disrupting the Purpose of Republicanism

Because social media amplifies and atomizes infighting, even the most dedicated and serious elected officials may find themselves in political jeopardy if they compromise or collaborate on the “wrong” issues. Representative government (republicanism) does not work well under 24-hour surveillance by the electorate—let alone an electorate animated by the Dunning-Kruger effect—or worse, professional trolls hired to attack the apostate the moment she steps out of line.

If one stammers at the upside-down world in which Liz Cheney is a “RINO,” social media made this alternate reality axiomatic by the same means that it became reasonable for so-called progressives to label President Obama a “warmonger.” The ordinary, even boring, job of governance has always operated behind the headlines of hot-topic issues. But due to the obligation to feed social media, nearly all politics are now performative, and the Member of Congress who does not entertain (i.e., does not deliver snappy comments on social media) may have a short and/or ineffective career.

No question that performance is always a part of politics, but social media enables more performative nonsense to flood the zone than was possible in the pre-digital era. Historically, a high-profile hearing, like a Senate confirmation hearing, would mainly be observed by the public through snippets and commentary edited by whichever news network we watched in the evenings. Meanwhile, low-profile hearings didn’t provide much opportunity to feed soundbites to constituents.

But speaking as someone who has watched a lot of back-burner hearings as part of his job, it is obvious that many are held for purely performative reasons because, of course, every Member has a social media person on staff who can make noise with a few provocative clips. That the substance of the hearing may be moot—or that most Americans won’t know it happened—doesn’t matter. The political value of the performance is enabled by social media, and this is a significant, and in my view negative, change to the nature of republicanism.

BlueSky is Nice but Can’t Fix the Problem

Yes, BlueSky is better (for now), and yes, I joined and started following people I believe to be thoughtful in their defenses of American democracy, etc. But the major adverse effects of social media cannot be eliminated either by design or the ethics of managers willing to keep hands off the algorithms. The affordances of the medium occur between the core technology and psychological responses to the experience—a thesis tragically supported by the fact that children can be coaxed into suicide by the material on their phones.

On that note, it will be interesting to see whether BlueSky supports platform accountability through legislation like the Kids Online Safety Act (KOSA), or a proper reading of Section 230, which was never meant to be blanket immunity for all providers hosting user content. Because ever since the so-called techlash of 2016, both the providers and their “digital rights” network have continued to push the narrative that somehow “restoring” the noble intent of the original internet should be the goal. But the original intent of the net’s first evangelists was not so noble (see Golumbia’s life work), and I think we all know Einstein’s definition of insanity.

Photo by: charlieblacker

TikTok Inspired Child Suicide Prompts a Sound Reading of Section 230

Last week, the Third Circuit Court of Appeals issued an opinion regarding Section 230 of the Communications Decency Act. It may be the strongest affirmation to date that the statute does not provide a blanket liability shield for all social platforms regardless of their conduct. Specifically, §230(c)(1) only immunizes platforms for liability that may arise from other parties’ speech, not from the platform’s own speech. And although the platforms have sought to argue that their “recommendation” algorithms, which push content to users, do not constitute speech, the courts aren’t buying it.

In the case Anderson v. TikTok, the appeals court reversed the lower court finding that the platform was automatically immunized against a liability claim involving the death of a child who attempted one of the many dangerous “challenges” that appear on social media. In this case, Nylah Anderson, age 10, died by accidentally hanging herself when she tried the “Blackout Challenge,” which dared people to asphyxiate themselves until they passed out. At issue for TikTok is not the challenge itself, started by an unknown third-party, but the “For You Page” algorithm which “recommended” the challenge to Anderson. Judge Matey, in a strident concurrence with the circuit court opinion, writes the following:

TikTok reads § 230…to permit casual indifference to the death of a ten-year-old girl. It is a position that has become popular among a host of purveyors of pornography, self-mutilation, and exploitation, one that smuggles constitutional conceptions of a “free trade in ideas” into a digital “cauldron of illicit loves” that leap and boil with no oversight, no accountability, no remedy.

Though the reference to St. Augustine implies a religious moralizing I might omit, Judge Matey’s accusation that social platforms host a “cauldron” of dangerous, illegal, and depraved material behind a veil of social good and constitutional rhetoric is indisputable. As a legal matter, had Anderson discovered the video challenge (e.g., via search), TikTok would likely be immunized by §230, but because a “recommendation” algorithm factored in the child’s conduct resulting in her death, this is an important distinction that could more clearly articulate a shift in judicial review of the statute and, we should hope, an overdue change in platform governance.

As Judge Matey further states in his concurrence, TikTok’s presumed immunity under §230 in this case is “…a view that has found support in a surprising number of judicial opinions dating from the early days of dial-up to the modern era of algorithms, advertising, and apps.” That view is properly dimming now, and by my reckoning, the Supreme Court will go where the Third Circuit went last week. In a pair of nearly identical cases, Gonzalez v. Google and Twitter v. Taamneh (2022), the plaintiffs, on behalf of victims of two ISIS-related terror attacks, sought to hold the platforms accountable for “recommending” ISIS recruiting videos. But because those claims relied substantially on meeting the standard for “aiding and abetting” under criminal law, the Court found little plausible claim for relief and, therefore, declined to address the question of §230 immunity.

But if Anderson (or a similar case) goes to the Supreme Court, I believe the justices will have little difficulty finding that a “recommendation” algorithm promoting a video challenge that led to a child’s death is a foundation for a liability case to proceed. As the Court stated in Taamneh, “When there is a direct nexus between the defendant’s acts and the tort, courts may more easily infer such culpable assistance.” In Anderson, with no other party acting as the direct cause of the child’s death, the facts are even simpler, revealing a clear nexus between the video challenge “recommended” by the platform and the accidental suicide. Further, this July, the Court held in the unanimous Moody v. NetChoice decision that social platforms “shape other parties’ expression into their own curated speech products.”[1] Under that rule, the Third Circuit finds that TikTok’s “recommendation” of the Blackout Challenge to Niyah Anderson plausibly constitutes the platform’s own speech, for which it may be held liable.

The reason I keep putting “recommended” in quotes is that at the time SCOTUS granted cert in the Taamneh and Gonzalez cases, I wrote a post opining that the courts, policymakers, et al. should take a jaundiced view of this too friendly term to describe an insidious function of social media. It is no longer controversial to say that platform operators manipulate what users see and hear, or that this manipulation can lead to disastrous results from disinformation campaigns in the political arena to drug-related deaths to suicide by little girls.

It is a familiar refrain that it takes a tragedy, or many tragedies, to change policy, and with the story of Nylah Anderson, and the many young victims she represents, we may finally see Big Tech’s hypocrisy on speech collapse under the weight of its own absurdity. The major platforms have played games with the First Amendment and §230 for nearly 20 years—conflating their business interests with users’ speech rights or asserting their own speech rights when necessary or asserting that nothing they do is their own speech—all depending on which potential liability the company seeks to avoid. Further, that confusion has not been helped in recent years by certain politicians who misstate the operation of the speech right to create political theater around allegations of bias.

Out of all that mess, it is notable that Justice Thomas, since at least 2020,[2] has restated the observation that online platforms will avail themselves of constitutional protection to engage in conduct like algorithmic “recommendation” but then invert the argument to shroud itself in the §230 shield. And then, the courts will stop a liability claim from even proceeding. As Congress, the Supreme Court, and now the Third Circuit have all reiterated, no industry in the country enjoys that kind of immunity, and perhaps this claim against TikTok will be the case that finally ends this unfounded and unreasonable privilege for online platforms.

[1] On a side note, this is reminiscent of the “selection and arrangement” doctrine in copyright law, which finds “expression” in the choices made by the author who engages in that conduct. All copyrightable expression is a form of speech.

[2] See dissent on the grant of certiorari in Malwarebytes v. Enigma.

Photo by:

Archives

Browse Topics