Lie Machines: How to Save Democracy from Troll Armies, Deceitful Robots, Junk News Operations, and Political Operatives
For the better part of four years, those sounding the alarm about the dangers of fake news and the perils of a post-truth world struggled to make the case that this was a matter of life and death. Try as they might to argue that a secure foundation of facts was the very basis of a liberal, democratic society—that such a society could not function without a common, agreed-upon basis of evidence—the concern seemed somehow abstract, intellectual, even elitist. Their angst was easily dismissed by their populist foes as the self-interested whine of a snobbish establishment. And then came the coronavirus.
When a pandemic is raging, it becomes harder to deny that rigorous, truthful information is a mortal necessity. No one need explain the risks of false information when one can point to, say, the likely consequences of Americans’ coming to believe they can deflect the virus by injecting themselves with bleach. (The fact that that advice came from the podium of the president of the United States is one we shall return to.) In Britain, Conservative ministers who once cheerfully brushed aside Brexit naysayers by declaring that the country had “had enough of experts” soon sought to reassure voters that they were “following the science.” In the first phase of the crisis, they rarely dared appear in public unless flanked by those they now gratefully referred to as experts.
So perhaps the moment is ripe for a trio of new books on disinformation. All three were written before the virus struck, before we saw people refuse to take life-saving action because they’d absorbed a baseless conspiracy theory linking Covid to, say, the towers that emit signals for 5G mobile phone coverage. But the pandemic might mean these books will now find a more receptive audience, one that has seen all too starkly that information is a resource essential for public health and well-being—and that our information supply is being deliberately, constantly, and severely contaminated.
The most vivid example remains the intervention by Russian intelligence in the US presidential election of 2016, in which 126 million Americans saw Facebook material generated and paid for by the Kremlin. But the phenomenon goes far wider. According to Philip N. Howard, professor of Internet studies at Oxford, no fewer than seventy governments have at their disposal dedicated social media misinformation teams, committed to the task of spreading lies or concealing truth. Sometimes these involve human beings, churning out tweets and posts aimed at a mainly domestic audience: China employs some two million people to write 448 million messages a year, while Vietnam has trained 10,000 students to pump out a pro-government line. Sometimes, it is automated accounts—bots—that are corralled into service. The previous Mexican president had 75,000 such accounts providing online applause for him and his policies (a tactic described by Thomas Rid in Active Measures as “the online equivalent of the laugh track in a studio-taped TV show”). In Russia itself, almost half of all conversation on Twitter is conducted by bots. Young activists for Britain’s Labour Party devised a bot that could talk leftist politics with strangers on Tinder.
Still, Howard writes in Lie Machines that the place where disinformation has spread widest and deepest is the US. He and his team at Oxford studied dozens of countries and concluded that the US had the “highest level of junk news circulation,” to the point that “during the presidential election of 2016 in the United States, there was a one-to-one ratio of junk news to professional news shared by voters over Twitter.”
It’s tempting to say that such material only has an impact at the margins, that only a relatively small number of people would ever be swayed by it. But the 2016 election was decided at the margins, the votes of fewer than 80,000 people in three swing states tipping the presidency to Donald Trump. In a 50–50 nation such as the US, a nudge to 51–49 is all it takes.
So these “lie machines”—consisting, Howard writes, of the governments or political campaigns that produce the lies alongside the social media platforms, algorithms, and bots that distribute them—matter gravely. They attack not just their specific target, such as Hillary Clinton in 2016, but what Rid calls the “liberal epistemic order, or a political system that places its trust in essential custodians of factual authority,” a category that includes science, the academy, journalism, public administration, and the justice system. For Rid, this is the order that in turn enables an open and liberal political order; “one cannot exist without the other.” Now that people can see the difference between a scientist warning of a coming pandemic and a demagogue implying that such warnings were a “hoax”—and now that they know the consequences of heeding one over the other—such arguments have gained a concreteness and urgency they might have lacked before.
How, then, should we define disinformation and how does it work? In Active Measures, the fullest, most elegant of these three books, Rid, a professor at Johns Hopkins University who grew up in what used to be West Germany, opens with an essay that clears up a few confusions. Foremost among them is the misconception that disinformation is necessarily false information. On the contrary, the hack-and-leak tactic, deployed to such potent effect against Clinton and the Democratic National Committee in 2016, worked only because it revealed information that was genuine. The leaked John Podesta e-mails really were e-mails sent to and from the chairman of Clinton’s presidential campaign. But as Rid notes, “even if no forgery was produced and no content altered, larger truths were often flanked by little lies, whether about the provenance of the data or the identity of the publisher.” So while Podesta’s risotto recipe was real, the hint by WikiLeaks that the e-mails had come from a DNC insider, possibly the young staffer Seth Rich, who was killed in a shooting incident in Washington, D.C., in July 2016, was not. (In his 2019 report, Robert Mueller went out of his way to dismiss the Rich theory as false, setting out how WikiLeaks had, in fact, been in touch with the Russian hackers who were the true source of the e-mail cache.)
All three books present accounts of that 2016 operation, which remains the definitive example, supremely instructive in the mechanics of disinformation. Ben Buchanan, who teaches at Georgetown’s School of Foreign Service, provides a helpful reminder in The Hacker and the State of the sheer diligence and seriousness of purpose exhibited by the Russians in their mission. The work began in 2014, possibly even earlier, as staff at the now infamous Internet Research Agency in St. Petersburg studied closely the ways Americans use social media, even traveling to the US several times that year to observe.
By 2016, IRA agents were posing as Americans online, making contact with political activists and organizers, assessing the lay of the land, concluding that they needed to focus their attention on “purple states,” a phrase they used internally. Next, they created hundreds of bogus social media accounts, crafting a persona for each one, complete with interests and hobbyhorses, always keeping a careful eye on the time zone inhabited by their fictitious alter egos. Just as call-center employees in Bangalore, working for UK companies, receive a regular digest of the plot twists of British soaps, enabling them to make apparently natural conversation with their customers, so the IRA’s trolls were supplied with a list of US public holidays, the better to pass as American citizens. The IRA rented servers inside the US and arranged relays so that their traffic appeared to originate on US soil.
But they did not work alone. They created groups on Facebook organized around the most divisive issues in American life: race, religion, identity. Buchanan provides examples: Secured Borders, Blacktivist, United Muslims of America, Army of Jesus, Heart of Texas—each one founded and administered by an agent of Vladimir Putin. Pretty soon, these groups were boasting hundreds of thousands of members. Some “were Russian operatives with fake accounts, but many were Americans who did not know they had fallen for a foreign influence campaign,” Buchanan writes.
The groups’ focus was unambiguous: to hurt Hillary Clinton. IRA managers told their staff to “use any opportunity to criticize Hillary and the rest (except Sanders and Trump—we support them).” They had productivity targets, so that if the number of anti-Clinton posts dropped off, the trolls were scolded, with a reminder that criticism of Clinton was nothing less than “imperative.” The output was either anti-Clinton or pro-Trump—or, in a third category usually aimed at minorities, some posts advised black Americans in particular that the choices were so awful that they would be better off not voting at all.
Some of the messages made the leap from Facebook to the campaign itself, with Trump surrogates and operatives picking them up and repeating them, unwittingly parroting themes originated in the Kremlin. To give things a further push, the Russians bought advertising on social media, including at least 3,500 ads on Facebook. It’s illegal, of course, to use foreign funds to influence a US election, but who was to know? The Russians had stolen the identities of several US citizens, so no one could spot that their ad buys were illegal. And, thanks to Facebook’s microtargeting algorithms, those ads reached exactly the right people: US voters passionate about whichever theme was being pushed, whether gun rights or abortion.
Not content with mere online influence, the IRA moved its destabilization-through-disinformation campaign from the screen to the streets. Russia’s Facebook pages convened rallies, hiring US citizens to stage political stunts. You might remember an image of an American dressed up as Clinton in a prison uniform, riding around in a cage on the back of a flatbed truck. Chances are high that you were looking at a stunt produced, directed, and funded by Russia. Worse, the Kremlin staged demonstrations and counter-demonstrations in the same place at the same time. “In one case,” reports Buchanan, “a Russian-run Facebook group planned a rally called ‘Save Islamic Knowledge’ in Houston while another Russian-run group organized the counterprotest: ‘Stop the Islamization of Texas.’ Police were deployed to keep the groups from physically clashing.”
The political logic here was not subtle, with the Kremlin identifying the fissures and fault lines of American life and driving a well-aimed digital wedge into each one. Russia wanted to elect Donald Trump but, perhaps above all, it wanted to intensify internal American rancor. Indeed, the former goal was, in part, a means to the end of the latter. Judged by that standard, it has been an extravagant success.
The natural impulse is to see Russia’s attack in 2016—and the one it is surely preparing for 2020—as a radically new feature of our hyperconnected world. Everything about it, all those bots and algorithms, seems novel. Yet Rid’s book is devoted to persuading us that it is in line with decades of history.
In rich detail, Rid walks us through a hundred years of political warfare, recounting the exploits powers both major and minor inflicted on one another via the disinformation units of their intelligence agencies. Some of the stories are hair-raising. We learn of Operation NEPTUN in 1964, in which Czech intelligence dispatched a team of underwater divers to Bohemia in the dead of night to drop four chests to the bottom of a lake, each one full of what purported to be Nazi documents. The boxes had been suitably treated to appear aged by twenty years of corrosion; inside were blank sheets of paper. The plan was for those to be replaced by authentic Nazi-era records supplied by the KGB from Moscow, where they had been held in state archives, along with “two or three forgeries” that would compromise several top officials in West Germany by apparently exposing them as onetime Nazis.
All went swimmingly. A Czech TV crew duly discovered the crates and hauled them to the surface, then handed them over to a team of unknowing government engineers who checked the boxes for explosives before surrendering the envelopes within, unopened, to an approved “group of experts.” That allowed the switch to happen, with the experts dropping in the stash of papers supplied by the KGB. The only problem was that Czech intelligence could never be sure that it hadn’t itself been played by its Soviet counterparts in disinformation: at one point, it suspected Moscow’s Service A might have forged all of the documents, though the Russians insisted they were genuine. Nevertheless, within a few months the Czech interior ministry was holding an international press conference trumpeting a haul of papers that reminded the world of the Nazis’ crimes and boosted opposition to West Germany throughout Western Europe. Mission accomplished.
The cold war was full of such antics, including a discreet and successful Stasi operation to engineer the first parliamentary vote of no confidence in the history of the West German republic, a feat pulled off not in public but by hoodwinking individual German politicians. The daring, the tradecraft, the stolen signatures and fake letterheads, the double- and triple-bluffs are hugely entertaining, at least from the safe distance of several decades, even if a few of the plots belong to the more outlandish, downmarket strain of spy fiction. The characters, though, are pure le Carré, not least Ladislav Bittman, the architect of the NEPTUN deception, who defected to the US and whom Rid meets in his home on Cape Ann, Massachusetts, where the old man stares out at the Atlantic Ocean, passing the hours of his retirement making modernist paintings. He could be Smiley, he could be Karla. In one fascinating passage, Rid muses:
It took a special kind of person to work in disinformation, on both sides of the Iron Curtain. Spotting weakness in adversarial societies, seeing cracks and fissures and political tensions, recognizing exploitable historical traumas, and then writing a forged pamphlet or letter or book—all of this required officers with unusual minds…free and unconventional thinkers, bookworms, writers, perceptive publicists with an ability to comprehend foreign cultures.
That “both sides” is important, because of course the Americans were in the disinformation business too, especially in the immediate postwar decade. Their methods involved not only well-known ploys such as the Congress of Cultural Freedom but also assorted other front organizations and publications, including, intriguingly, jazz and astrology magazines aimed at the East German market.
The objectives for the two sides were, true to the spirit of le Carré, the same. Just as Moscow sought to undermine the image and self-confidence of the West, so the West, and the US in particular, sought to do the same to Moscow. But what Rid discovers is that while Russia kept going right until the bitter end, “the West deescalated” its disinformation hostilities following the construction of the Berlin Wall in 1961. Rid doesn’t offer much by way of explanation, leaving the reader to suspect that Western spymasters concluded that there was no active measure they could concoct that would better alienate citizens of the Eastern bloc from their masters than de facto imprisonment behind a high wall topped with barbed wire. (Count that as one more reason to doubt the rumors, now the subject of a hit podcast, that the Scorpions’ post–Berlin Wall hit, “Wind of Change,” was a CIA job.)
Not that “active measures” were ever solely a cold war phenomenon. Until 2016, the greatest-ever act of foreign electoral meddling was one committed against the United States not by Moscow but by London. Buchanan recalls Britain’s efforts to draw the US into the war against Nazi Germany, efforts that did not rely solely on the rhetorical gifts of Winston Churchill. Before the Republican convention in 1940, for example, delegates seemed in a mood to nominate an antiwar, isolationist candidate to take on Franklin Roosevelt. But opinion shifted after the publication of a poll, which surprisingly showed that three in five convention delegates backed Britain in its struggle against Hitler. That helped the former Democrat Wendell Willkie to win the GOP nomination, from which perch he offered no opposition to Roosevelt’s transfer of American destroyers to the Royal Navy and kindly lost the election to FDR, both of which outcomes delighted London.
But here’s the thing: that poll never existed. It was one of multiple exploits by a team led by William Stephenson—later immortalized in A Man Called Intrepid by William Stevenson—who cooked up stories galore to discredit the isolationists and boost the case for war among the US public. In case the parallel with 2016, operational if not moral, isn’t clear, Buchanan writes:
Here was direct interference in United States presidential politics by a foreign actor, aided by the spread of false information, the manipulation of popular media, the clever timing of leaks and lies, and the creation of propaganda that aligned with preexisting narratives.
In other words, active measures were not invented in twenty-first-century Russia. They were such a routine feature of the last century that the US and the Soviet Union, Buchanan estimates, meddled in more than one hundred elections in other countries.
There might be some comfort in that, as if the current assault on facts and truth were merely the latest iteration of a threat we have lived with for decades and which we can, demonstrably, survive. That, though, depends on the answer to a tricky question: Is today’s disinformation merely different in degree from that of the past, or different in kind?
The continuities are clear enough. The longevity of Russia’s commitment to active measures is striking. Rid begins his book with a delicious tale of early Bolshevik intrigue, in which a White Russian aristocrat was turned and used to feed false comfort to his fellow tsarists, assuring them there was no need to take action because the Communist regime would soon collapse from within. In Rid’s account, Moscow’s pursuit of active measures continued even after the Soviet Union crumbled into dust. The end of the cold war did not mean the end of hostilities. It was, writes Rid, no more than a “temporary setback for the art and craft of disinformation.” Those engaged in it were cynically amoral then, and they’re cynically amoral now.
And yet it would not be right to conclude that today’s disinformation efforts are simply a high-tech version of those of the past. The differences are more substantial than that. Today’s active measures are simultaneously more personal and much broader in reach than before. While KGB operatives in the 1950s might have placed a forged pamphlet or bogus magazine in front of a few thousand readers, their heirs can now microtarget millions of individuals at once, each one receiving bespoke messaging, designed to press their most intimately neuralgic spots. Those engaged in what Howard calls “computational propaganda” don’t merely mine the attitudes you’ve expressed on social media; they can also draw conclusions from your behavior, as recorded by your credit card data. What’s more, think of all the data gathered by the connected objects around you—the Internet of things—monitoring your sleep, your meals, your habits, your every move. This reveals more about you than your browsers ever could, says Howard, adding, arrestingly, that we’ve been “focusing on the wrong internet.”
It’s this blend of “massive distribution, combined with sophisticated targeting” that is new. The work is so much easier too, requiring little of the fine, almost artistic skill demanded of the master forgers and tricksters of yore. In the earlier era, only governments, through their intelligence agencies, had the money and muscle to attempt such work. Now the cost of production is low, and so is the bar to entry.
What’s more, technological advances promise to make disinformation easier still and more effective. It’s already possible to create fake audio and video; it can’t be long before fake fact-checking sites follow. Chatbots are in their infancy, but they are growing more sophisticated. The future may see not only your Twitter feed dotted with AI bots, but even your WhatsApp messages, filled with “digital personalities” engineered to look and sound like people you know.
The heart of the matter is data, the resource that makes all this possible. For Howard, junk news is merely the symptom; the disease is the “monopolization of information” in the hands of a few tech giants. It used to be the churches that held the important information about us, he writes: our births, deaths, and marriages. Then it was governments and libraries. “Now a handful of technology firms have the best data on us as individuals, on our networks, and on public life,” and they sell both that information and the tools to exploit it to anyone willing to pay.
There’s a last difference in kind from the political warfare of the past, though none of these authors addresses it directly. Put simply, there can never have been a world leader so willing to amplify and echo the hostile messages of his most devoted adversary as Donald Trump. Only the most optimistic Kremlin spymaster would ever have dreamed of a US president who himself, unbidden, encourages the American people to lose all faith in their institutions, to distrust their media, scientists, judges, and intelligence agencies, even to take wild risks with their own health and so make a vicious pandemic worse. There is surely little need for active measures—spreading conspiracy theories or promoting bogus remedies—when the man in the Oval Office will do that work for you.
What, then, can be done to arm ourselves against the next decades of informational war? There are some mechanical steps worth taking, which sound almost too basic to spell out. One can only admire Mitt Romney’s 2012 presidential campaign, which, alert to the threat of foreign hackers and their interest in his choice of running mate, devised code names for the potential candidates and communicated only via computers unhooked from the Internet. US election officials at the federal and state levels would be wise to regard 2016 as a trial run for the mayhem Moscow might be plotting for 2020, viewing the various attacks on voting systems four years ago as, in the words of Franklin Foer in The Atlantic, “casing the joint.” Some rudimentary electronic defenses are missing and need to be put in place.
That is especially true given the nature of the incumbent. It is surely not wise to assume that Trump would take defeat gracefully, quietly packing his bags and waving farewell from the South Lawn. Trump is bound to claim that the vote was rigged, that the ballots were unsafe, and that the result in battleground states was void. With that in mind, the sage election official will look to ensure a verifiable paper record of all the votes cast. Not that such a precaution would restrain a president so determined to cling to office that, as some fear, he would invoke emergency national security powers, claiming a foreign adversary—say, China—had meddled in the election. In that scenario, a confected Department of Justice investigation into foreign intrusion might just offer a way for Trump to swerve around the electoral college and throw the election to the House of Representatives where, because the vote would be by state delegation, with one vote per state, Trump would be likely to win. In such a situation, a documented record of votes cast on November 3 would at least be a powerful exhibit in the court of public opinion. To ensure such a record, the most obvious mechanism is not so much low-tech as no-tech: a British-style ballot paper marked by a simple cross, with the papers counted by hand. No voting machines, no hacking.
Failing that, mail-in ballots would not only present an obvious remedy to the conundrum of holding an election in the era of social distancing, they’d also promise a measure of protection against a repeat Russian effort to swing the 2020 election: mailed votes automatically provide their own verifiable paper record. Hackable machines would still have to count them, of course—and a committed election-wrecker could always try to ensure that some ballots get “lost” en route or, no less damagingly, claim that they had—but for all Trump’s drum-banging about the risk of fraud, absentee ballots do at least offer the basic safeguard of a documentary record of a voter’s choice. It’s wearily predictable that a president who has never taken the threat of Russian interference seriously—who indeed is affronted by the mere mention of it in his presence—opposes even the modest precaution of absentee ballots.
Perhaps this debate has come too late. There are alarming signs that election supervisors across the US haven’t left enough time to protect themselves—a situation not helped by Senate Republicans’ refusal to pass a bill that would have afforded some protection against a Moscow offensive, replacing it with legislation that funded new voting machines but did not insist on security measures. In truth, and more broadly, if US elections are to be regarded as safe, they need to be put on a radically different legal footing—one that would overturn the Citizens United judgment that allows the funding of political campaigns to be so easily kept mysterious.
Howard offers a five-point manifesto, aimed chiefly at big tech’s monopoly on data. Some of his demands are innovative, including citizens’ right to donate their own data to favored political organizations, so that those players can begin to compete on something like level terms with the tech giants and those who currently pay to use their services. He also advocates “mandatory reporting on the ultimate beneficiaries of data,” much as arms manufacturers can be compelled to reveal the end-users of their products, and a tithing system, whereby 10 percent of ads on social media platforms are given over to public service announcements. Data is power, and Howard demands that we share it.
Politicians obviously need to be more alive to this menace—it’s grim to recall Barack Obama’s feeble response to the Russian attack in 2016, merely telling Putin to “cut it out”—but so do all those who write about and analyze politics. Clearly, every time a journalist wrote a story about the hacked DNC emails, they were doing Russia’s bidding, but the problem is bigger than that. Buchanan is right to suggest that “while most policymakers and scholars understand what nuclear weapons and tanks do, the possibilities, pitfalls, and processes of hacking missions are comparatively opaque.” Information warfare is designed to bamboozle, but its digital variant can be especially baffling to the nonspecialist.
Nevertheless, the only true protection against active measures, whether by Russia or anyone else, is to deny them the openings they rely on. Those 2016 attacks were devilishly ingenious, driving wedge after wedge into America’s most seismic fractures, but none would have worked had those divisions not been there, ready to exploit. A democracy such as the United States will always be divided—of course it will. But Americans’ best defense against foreign enemies might be to stop seeing political opponents as domestic enemies. Russia’s exploits work because Americans are too quick to turn viciously against one another. The culture war has made the country vulnerable in the disinformation wars. Working for a truce in the one might be the best hope for victory in the other.