The American foreign-policy community has spent the past year arguing about how to counter Russian influence operations. Newspapers such as the New York Times have written article after article describing how Russian trolls ran campaigns to discredit Democratic nominee Hillary Clinton’s presidential campaign. Twitter commenters with hundreds of thousands of followers claim that Russian influencers and agents-provocateurs are everywhere on social media. Research centers warn that Russia has the “motive and the means” to “undermine democratic political and social institutions,” suggesting that Russia is capable of destroying America’s democracy in very short order.
Meanwhile, elite bipartisan initiatives such as the Alliance for Securing Democracy have charged Russia with engaging in a coherent and tightly organized campaign across social media platforms, “pulling strings on both sides of the Atlantic,” undermining democracy in the United States and Western Europe. They propose that the United States counter Russia by developing Cold War-style strategies to “defend against, deter and raise the costs on Russian and other state actors’ efforts to undermine democracy and democratic institutions.”
All these people are not only getting Russian influence operations wrong. In an important sense, they are also amplifying the effects of the very operations they decry.
For sure, Russian influence operations are real, but they are neither as Machiavellian nor as successful in changing people’s minds as they first appear. As experts on Russia such as Julia Ioffe and Masha Gessen have argued, Russia is not working according to a master plan carefully laid-out laid out by President Vladimir Putin. Instead, a loose collective of Russians, with incredibly meager resources, have been working together in a disorganized way to probe American democracy for weaknesses. Instead of persuading people to vote for Donald Trump, and against Clinton, they have wanted to create chaos and paranoia — and they have succeeded in stirring confusion only because there were so many weaknesses for them to exploit in the first place.
A Cold War response of punishing and deterring Russia may feel satisfying, and it is probably worth carrying out whenever presidential politics allow it. However, punishing Russia is mostly irrelevant to the underlying problem. If a semi-incompetent social media campaign is all that one needs to send American politics into a halting state, then America’s troubles are far more fundamental than Russian interference. Indeed, efforts to combat Russian influence operations, unless they are very carefully designed, may make the problem worse. Opponents of Russia, who feed public fears that Russian influence is everywhere, may be working as inadvertent force multipliers of paranoia.
If it is as easy to compromise democracy as Americans now fear, punishment-based approaches are likely to end up as gigantic games of whack-a-mole. Instead of looking to punish attackers, policymakers need to start thinking about strengthening the system of democracy itself.
Here, the technical language of operational cybersecurity, rather than international relations, may prove most useful. Cybersecurity provides two key tools for thinking about a problem — “threat modeling” and the concept of “attack surface.” Rather than focusing on specific states, a threat model allows us to think clearly about the broad variety of actors who might attack American democracy. Analyzing attack surface, meanwhile, allows us to think about the “vectors” that these actors might use – that is, the vulnerabilities they might exploit — to weaken democracy, and how to limit or close them off.
Together, these tools allow us to understand the real threat. There are many actors besides Russia who could target the weaknesses of American democracy, including actors within America itself. The attack surface of American democracy is extraordinarily broad and provides many opportunities for manipulation. If America is to shore up those vulnerabilities, it will have to fundamentally change how it thinks about democracy and security.
It is not surprising that Russia is trying to use social media against the United States. The real puzzle is why these operations are succeeding so well.
As soon as Russian leaders figured out a decade ago how to counteract online agitating and organizing in their own country by, to quote the New Yorker’s Adrian Chen, “seeding doubt and paranoia, and destroying the possibility of using the Internet as a democratic space,” they began to weaponize these techniques for use against populations abroad, spreading confusion in Ukraine, the Baltic republics, the Nordic democracies, France, Germany, and, eventually, the United States. The evidence suggests, however, that Russia’s operations have always been poorly organized and opportunistic. Sometimes, they are laughably inept. Fake posts developed for use in the United States, for instance, are often written in bad English. Russia’s reported spending on Facebook advertising was a pitiful trickle compared to the torrents of cash spent by both the Trump and Clinton campaigns to influence voters.
Russian online influence operations nonetheless seem to be working better against the United States than other countries. Research suggests that Russia’s “MacronLeaks” operation was far more successful in attracting the attention of English-speaking alt-right activists than French voters. Germany, too, seems to have been better able than the United States to shrug off efforts to shape its political conversation.
Russia’s relative success in the United States is not thanks to the unique strategic insight of Putin. It is because Russian operatives have chanced upon real weaknesses in U.S. democracy, and American elites are unintentionally giving them a helping hand. While France and Germany have their own social divisions, they do not face the specific problems that America faces.
In America, more than in most other Western countries, there is a basic failure of democratic knowledge. In a well-functioning democracy, citizens agree broadly on facts and have some trust in the democratic system, allowing democracy to harness different perspectives and put them to good use. In America, in contrast, distrust and profound disagreements over facts have led to a kind of crisis of democratic knowledge that leaves democracy open to outside manipulation.
Over the last two decades, the common knowledge of American democracy has been undermined. As Alexis de Tocqueville warned could happen, the structures of shared knowledge are being weakened by democratic politics itself. Politicians — especially on the right — have cast doubt on sources of authority such as science and government, telling their supporters that they shouldn’t trust experts. Finally, the public itself, on its own initiative, has become less trusting of traditional institutions such as the Roman Catholic Church as they have revealed their feet of clay.
As a result, those who were disengaged from politics have become more so, while those who are engaged have become ever more partisan. The result is that most people don’t care about politics, and those who do are likely to have radically different understandings of the challenges faced by America.
Disaffection can be healthy up to a point. Many traditional institutions have failed badly and do not deserve people’s trust. Social movements such as Black Lives Matter have uncovered gross problems in how American democracy actually works. Moreover, strong partisanship can have real political benefits under the right circumstances. It gives people with sharply different values the voice to express them and the incentive to come up with good ideas to compete for votes.
Yet when people with different perspectives stop sharing a common basis of knowledge, democracy is liable to pull itself apart. Parties become enemies rather than competitors. When people stop trusting any institutions, they are likely instead to start thinking that the democratic process is rigged, and to pin their hopes instead on cranks and conspiracy mongers.
This leads to big problems in the age of social media. Platforms such as Facebook and Twitter provide fertile ground for conspiracy theories and make it even easier for people to build their own private universes of fact. On one side, rumormongers like Mike Cernovich tell Trump supporters (his fans include Donald Trump Jr.) that Hillary Clinton was involved in a pedophile ring run from the basement of a popular Washington, D.C., pizzeria. On the other, Twitter personality Louise Mensch reassures hundreds of thousands of diehard Trump opponents that deep-state institutions have already indicted the president and are clearing the way for Sen. Orrin Hatch (R-Utah) to take over.
Social media platforms are also more open to deliberate manipulation. Academic research suggests that as much as 20 percent of all Twitter accounts are fake. Fake accounts on Twitter and Facebook can start new rumors. They can make it seem as though these rumors are widely believed by broadcasting and rebroadcasting them via multitudes of “bot” accounts, which are linked to automated algorithms that relentlessly push out content. People on Twitter are as likely to retweet messages that are being shared by bots as messages shared by real people.
The result is that social media increasingly resembles a science fiction dystopia in which nothing is real anymore. Fake stories mutate and proliferate through people’s social networks like an aggressive carcinoma. Fabricated or hijacked social media accounts play a crucial role in spreading these stories and making it seem as if multitudes of others believe them. Just as in the science fiction stories of Philip K. Dick and Frederik Pohl, people can’t tell whether the strangers they are talking to and listening to on social media are real people or machine intelligences pretending to be human.
Again, Russia has not engaged in an organized, planned campaign, so much as an opportunistic and quasi-coordinated set of probing attacks, conducted with relatively limited resources. Adapting an approach that was honed against its own population, Russia has used online identities, farmed out networks of bots, and used them to relentlessly pump out social media messages to spread confusion and distrust.
In some countries, these attacks have had minimal success — creating a political nuisance, but nothing more. In others, such as the United States, they have had greater consequences. This is not because Russian efforts in the U.S. election were crafted with precision, but because American democracy is in such trouble that even relatively inept efforts can succeed.
The United States is facing something close to a perfect storm, where badly organized, badly executed, and badly funded outside influence operations have had outsized consequences. Russian social media activities have probably somewhat enhanced the effectiveness of other Russian activities, such as the presumed hacking of Democratic Party computers, and may have spread initial confusion. However, their biggest consequence today may be to spread the belief that Russian influence operations are ubiquitous and wildly successful.
It is unsurprising that there is little evidence that Russian activities have, for example, led to significant changes in what people think. While social media surely influences what people do, think, and say, it is hard to use it effectively on the cheap. Russian media accounts were only one relatively small grouping among a vast crowd of would-be influencers, each trying to shout down the others. Most of Russia’s efforts likely simply disappeared in the noise.
Yet the simple fact of their existence has spread suspicion and even paranoia about their extent and effectiveness. Even during the elections, people were claiming that apparently genuine bursts of online political activity were actually the product of Russian disinformation campaigns. Since the election’s conclusion, speculation has exploded, especially among dismayed Clinton supporters, that Russian social media operations played a key role in getting Trump elected, and that many of their apparent online opponents are in fact paid trolls for the Kremlin.
Such fears are exacerbated because social media conversation takes place on privately owned information systems, whose owners do not disclose everything they know. It is hard to know the extent to which Twitter has been compromised. While Twitter claims that independent estimates grossly overstate the number of fake accounts, it refuses to release its own data. It is nearly impossible for anyone apart from Facebook itself, or researchers who have signed extensive nondisclosure agreements, to know exactly what is happening on Facebook’s social networks.
Facebook uses proprietary algorithms to determine which stories to emphasize in users’ news feeds, and which to bury, and makes it impossible to know how stories are spreading across its network as a whole. This lack of visibility makes it easy for the public to fear that Russian influence is spreading through a dark realm of social networks that are invisible to those who are not directly involved in them.
In principle, efforts to combat Russian propaganda could help allay these dark fears. However, in practice they are likely making them worse. By exaggerating the actual consequences of foreign influence operations, American elites are further undermining the confidence and shared knowledge that American democracy needs to function. They are tacitly encouraging Americans on the liberal left to build their own private universe of facts, in which Russian influence has pervasive political consequences.
Of course, Trump and his supporters have already built an alternative reality on the right, in which America is besieged by an alliance of Muslims and liberals, who stole the popular vote by allowing millions of undocumented immigrants to vote illegally. A democracy in which both the liberal-left and the right have drifted into their own self-reinforcing mythologies will be incapable of defending the common interests of its public, either at home or abroad.
Trump’s relationship with Russia means that it will likely be left to a future administration to punish Russian influence operations. In any event, the more important answers to America’s problems lie at home, where social media and elite panic have accelerated the decline of the country’s basic knowledge structures.
For sure, both past history and current politics suggest that Russia may be particularly well positioned to take advantage of weaponized paranoia. Few other actors’ manipulations are likely to be amplified so much by domestic paranoia about manipulation and Manchurian candidate strategies. However, the threat model that America faces is not further Russian attacks, or even necessarily further attacks by nation-state adversaries.
Now that the weaknesses of American democracy have been revealed, other actors — including some much closer to home than Russia — can look to increase democratic disarray for their benefit. The more likely threat that America faces is of a variety of smaller attacks on specific aspects of democratic knowledge, each aimed at particular ends, but accumulating to create massive overall damage. For example, there is nothing stopping American or foreign billionaires from undertaking further targeted attacks against aspects of public knowledge that they find uncongenial. Social media tactics could dramatically amplify attacks on scientific knowledge about global warming or efforts to undermine faith in the democratic process.
This threat model is likely to become much worse in the near future. These Russian efforts were clumsy, ill coordinated, and technically unsophisticated. The next attacks — whether they come from Russia or elsewhere, are likely to be far more technologically adept. As danah boyd, a prominent expert on social media, has argued, future attacks will likely use techniques designed to manipulate the machine learning algorithms that allow social media platforms to function, as well as carrying out their own machine learning attacks.
The attack surface is very hard to defend. Many of the underlying flaws of American democracy will take decades of work to remedy. A real cure cannot be imposed from above, since it requires a root and branch reorganization of American politics to make it more competitive, such that political parties have to fight to persuade voters rather than carving up the country into segregated ideological safe spaces. That would make it easier to build up healthier structures of knowledge, based on implicit agreement about which problems America faces, and explicit arguments over how best to solve them.
In the shorter term, policymakers and policy thinkers need to focus on more specific ways in which existing knowledge structures can become vectors of attack. Most obviously, social media makes it much easier to weaponize confusion and paranoia. U.S. regulators and social media companies such as Facebook need to make some hard choices and make them soon. More and more of America’s public conversation is happening in privately owned spaces, where no one knows who is doing what to influence whom. Facebook’s and Twitter’s moves toward greater disclosure and greater supervision of who buys ads are welcome, but not nearly sufficient. It will be hard for people to believe that American democracy is genuinely fair and transparent as long as much of its activity stays hidden behind the firewalls of online social networking companies. Nor will Facebook’s efforts to segregate political news from people’s personal feeds necessarily help. Without careful design, it may just mean that rumors flow more quickly and effectively among the political junkies who are most likely to engage with bad information.
Yet, the attack surface does not end with social media. Other forms of democratic knowledge creation now have direct implications for national security. For example, if the 2020 census process is damaged through a combination of social media engineering and partisan self-interest, U.S. democracy will be compromised, both because many people will not be represented and because people will disagree bitterly over whether the process was fair. Partisan efforts to damage confidence in voting, based on bogus statistics, could have serious repercussions if there is a contested result in the 2018 congressional elections or the 2020 congressional and presidential vote.
Under a nightmare scenario, outside hacking could combine with partisan efforts to spread fear about the U.S. voting system to create disarray. The U.S. intelligence community has formally found that Russian hackers apparently tried to gain access to state electoral registries. The Russians probably did not want to secretly bias voting so much as to create a controversy over whether manipulation had happened, as they had previously attempted in Ukraine. If outside actors succeeded in this kind of attack in the future, it could serve as a massive force multiplier for partisan disagreement, leading, even in an optimistic scenario, to years of political chaos. As a group of experts has concluded: “Simply put, the attacker might not care who wins; the losing side’s belief that the election was stolen from them may be equally, if not more, valuable.”
If America is not to find its democracy systematically dismantled, it needs to strengthen its structures of shared knowledge and trust in democracy as a matter of urgent priority. This most obviously involves strengthening voting systems and registries (which are now run through a mishmash of badly secured systems at the local and state level) against attack. Strengthening the census, rebuilding confidence in voting, and re-establishing knowledge structures that help partisans work in harness are usually thought of as exercises in civics. They now have crucial implications for national security, too.
This time, Russia probably lucked out. In the future, malicious actors will be able to use far more sophisticated knowledge weapons, in a country ever more shaped by Google’s and Facebook’s combination of machine intelligence and social feedback. It will be hard to think about, let alone confront these problems, if U.S. policy elites continue to allow Cold War fantasies to distract them from the fundamental weaknesses of American democracy.
Henry Farrell is a professor of political science and international affairs at George Washington University.