Pundits have worried that technology might doom democracy since Donald Trump’s election as U.S. president in 2016. It’s true that social media can benefit aspiring autocrats. Populists in particular latch on to social media today as a way to connect directly with people, bypassing restraints on their behavior that political parties would have provided in the pre-internet age. They can also profit from echo chambers, which reinforce the sense that a whole people uniformly supports a populist leader.

Yet social media is not inherently populist. And if populists do well this year, it will not be because there are no tools or strategies to stop them.

To combat populism, democracies need political will. They must not only push for better platform design and regulation but also work to strengthen what some consider a thoroughly old-fashioned institution: political parties that have the capacity to rein in leaders threatening democracy.

Álvaro Bernis illustration for Foreign Policy

At first, social media was greeted with great optimism. In what now feels like a different era, promoters of democracy looked to Twitter (now known as X) and Facebook as tools to help uprisings against autocrats everywhere. But just as the Arab Spring turned to Arab Winter, enthusiasm morphed into pessimism. Panic ensued in 2016, after the double shock of Brexit and Trump’s election. Liberal commentators were quick to identify what they saw as a major culprit of the world’s twin populist disasters: social media and, in particular, echo chambers. Not only did liberals veer from cheering to jeering. They also indulged in nostalgia for a supposedly golden age of responsible gatekeeping by journalists. The wild swings in opinion and the idealization of the past were signs that we have yet to find our bearings when making sense of new media.

Social scientists today know a bit more than they did in 2016: Filter bubbles—or online echo chambers curated by algorithms—exist but are much less common than often assumed; they are not the main cause of polarization, even as they help spread disinformation and propaganda more swiftly; and our offline life is in many ways less diverse than our online existence.

What makes social media unique is that it allows for what can seem like a direct connection between political leaders and potential followers. This is particularly useful for populists, who claim that only they can represent what they often call the “real people.” This implies that all other contenders for power do not represent the people, since, as the usual charge goes, they are corrupt. It also implies that some citizens are not part of the “real people” at all. Think of Trump complaining that his critics are not just wrong about policy but that they are “un-American” or even—as he put it at a Veterans Day rally last year—“vermin.” The point of populism, then, is not just criticism of elites. After all, finding fault with the powerful is often justified. Instead, the point is to exclude people from the people: other politicians at the level of party politics and entire groups—usually already vulnerable ones, such as Muslims in India—at the level of the citizenry.

This seemingly direct connection contributes to the erosion of political parties. Populism is about denying and, eventually, destroying pluralism; well-functioning parties can push back against this and rein in populist political entrepreneurs. Some countries even require parties by law to have internal democratic structures. (The radical right-wing Dutch populist Geert Wilders’s Party for Freedom, which won the most seats in last November’s elections, would not be allowed in those countries because Wilders is the only official member.) Of course, parties unite partisans. But partisans often disagree on how principles they share should translate into policy. There is nothing strange about parties forming legitimate opposition to their leadership, and it has often proved crucial in providing a check on leaders. There’s a reason that populists such as Indian Prime Minister Narendra Modi and Hungarian Prime Minister Viktor Orban run their parties in a highly autocratic fashion.

Filter bubbles can help populists sell their core product: the notion of a homogeneous people united behind the populist leader.

To be sure, the sense of directness created by social media is an illusion. Social media mediates, after all. Yet the prospect of an unfiltered encounter—however misguided—promises authenticity and a sense of connection that was once available only at exceptional moments, such as at a party meeting or mass rally. The political theorist Nadia Urbinati has suggested the paradoxical-sounding term “direct representation” for this relationship: Anyone standing between citizens and their representatives seems to have disappeared.

The work of getting people to the polls used to be done differently. As the political scientist Paul D. Kenny explains in his book Why Populism?, before the age of social media, mobilization depended on clientelism or a well-organized (put more bluntly: highly bureaucratized) political party. Parties and candidates promised supporters material benefits or bureaucratic favors in exchange for votes. This was costly, and costs would rise steeply if political competition intensified or more power brokers entered the fray. Bureaucratic parties are also expensive to maintain. Party officers have to be paid, even if they can count on volunteer work from idealists who sacrifice their weekends to distribute leaflets or canvass door-to-door.

As Kenny points out, social media cuts the costs of mobilization, especially for celebrity candidates such as Trump, who can draw on their pop culture credit. In the old days, when print and TV were dominant, propaganda feedback loops would have been constructed at great costs by party strategists; today, they are created for free by companies that want to maximize engagement for the sake of profit.

As with influencers, a politician’s online presence requires constant curation, so it is not entirely costless. Trump might have written his own tweets, spelling mistakes and all, but others need to pay tech-savvy teams. Social media might work best for those who already treat parties as instruments for marketing a personality rather than developing policy. Take former Italian Prime Minister Silvio Berlusconi, whose public relations specialists created the Forza Italia party for him in the 1990s and organized it like a fusion of soccer fan club and business enterprise. It is not an accident that Berlusconi joined TikTok before the most recent Italian elections in 2022 (even if the ragazzi he tried to appeal to might have found his performance, as young adults would say, cringe).

The most successful politicians can tap into both forms of support. For instance, Modi, with his enormous cult of personality, has emerged from a mass membership party with a bureaucratic apparatus and can rely on the free labor of partisan foot soldiers. Yet he has also built a following online, where he has been able to present himself as a celebrity above party politics.

Once populist leaders establish the illusion of direct connection, they find it easier to discredit traditional mediators, such as professional journalists, by claiming that they distort politicians’ messages. That can translate into fewer pluralistic debates and fewer opportunities for reporters to ask inconvenient questions. Modi and Orban have not held a genuine press conference in many years; Trump and Israeli Prime Minister Benjamin Netanyahu have declined to join debates before elections. Trump’s refusal to appear onstage with current Republican candidates might seem like a risky gamble: As candidate Ron DeSantis has tried to point out, the front-runner seems afraid to engage the rest of the pack; plus, he’s losing an opportunity to fully display his knack for the memorable put-down. But Trump is following the autocrat’s playbook: to appear above the fray and portray yourself as the unique embodiment of the popular will. Why stoop to the level of the competition if you’ve already told your supporters that everyone else is corrupt or, at the least, completely unrepresentative of their views?

Filter bubbles can therefore help populists sell their core product: the notion of a homogeneous people united behind the populist leader. Algorithmic curation designed to increase engagement with like-minded users amplifies this dynamic. Platforms often suggest what to watch or click on next. Anyone looking up Orban on X, for instance, will likely find an assortment of far-right content. When I recently checked his account, I was shown tweets from the Russian foreign ministry and U.S. presidential candidate and conspiracy theorist Robert F. Kennedy Jr.

To be sure, these online bubbles do not form in a vacuum. In the United States, plenty of people do live in a far-right bubble, without any contact even with center-right outlets such as the Wall Street Journal. This bubble is not the result of Facebook or X, however. As social scientists at Harvard University demonstrated in a 2018 study, its contours were shaped by the enormous success of right-wing cable news and talk radio in the 1990s. Social media just came on top of that infrastructure. If social media itself made for a world where conspiracy theories and hate always reign, we would see the same outcome in every country—but we don’t.

Álvaro Bernis illustration for Foreign Policy

As Elon Musk’s changes at Twitter have demonstrated, those who control platforms and their underlying machinery can manipulate online discourse. Since he took over the platform in 2022, Musk has not only arbitrarily suspended journalists but also weakened the rules—and reduced staffing—for content moderation. As Musk has replatformed white supremacists and other hatemongers, minorities such as transgender people have become less protected.

In halfway-functioning democracies, capricious oligarchs such as Musk get to govern platforms almost singlehandedly. In countries on the path to autocracy, the state itself can successfully pressure platforms to do its bidding, as India has done with Twitter by forcing it to block politicians, activists, and even the BBC. In outright autocracies, governments are perfecting what the social scientist Margaret Roberts has called friction and flooding. Rather than simply rely on fear created by widespread repression, as traditional dictatorships would, autocracies now “flood” the web with information to distract users and use intentional technical glitches (“friction”) to make it more difficult for citizens to access certain sites. These regimes know that censorship can draw attention to scandalous content; the truly savvy make it disappear. Such techniques are ubiquitous in China, as is surveillance. Aspiring autocrats, including right-wing populists vying for power in democracies, will no doubt try to copy this repertoire.

One problem with social media in its current form is that it gives too much power to only a few people.

To be sure, populists cannot be prevented from building their own counter-publics online, just as parties cannot—and should not—be hindered as they bring together followers. Freedom to assemble and associate means that like-minded people have every right to get together with others who share the same commitments. One would not want authorities to start shutting down safe spaces for groups devoted to empowering minorities, for instance, just because they happen to be insufficiently pluralistic. Ideas to combat online homogeneity through injecting viewpoint diversity into online life are well intentioned but impractical. The jurist Cass Sunstein, for example, has suggested a “serendipity button,” which could very well come out as, “Now that you’re looking at the feminist viewpoint, how about clicking on the anti-feminist one?”

A more nuanced view of online political life does not mean that democracies must tolerate the incitement of hatred. Platform design makes a difference: As the political scientist Jennifer Forestal has shown, Reddit, for instance, makes for a more diverse conversation than Facebook Groups. Reddit allows for communities to form but keeps borders between subreddits permeable; it also empowers both moderators and users to stick to rules agreed on by an online community.

Content moderation in particular should be mandatory, as it is in Germany, rather than a luxury that a platform controller such as Musk has the power to dispense. Moderation can be abused, but that is the case with any attempt to control media power. (Libel laws can be—and are—exploited by undemocratic actors, but that does not mean we should dispense with them altogether.) To forestall this, content moderation must be as transparent as possible and subject to proper oversight; the “black boxes” of algorithms should be opened at least to researchers so that they can help policymakers understand how social media platforms are run. This might sound like a pipedream. But the European Union has been pursuing these goals with its recent Digital Services Act and Digital Markets Act, which so far have prevented Facebook from launching its X clone, Threads, in the bloc due to its failure to comply with privacy regulations.

Legislation and education will be important tools for democracies. The business models of social media, which are based on maximizing engagement through offering ever more extremist content, are not beyond political regulation. Democracies should also invest serious resources in teaching media literacy—something that many leaders affirm in the abstract but that, just like civic education, always gets short shrift in the end, since “hard” subjects such as math are seen as more important for global economic competition. Not least, democracies must not treat social media in isolation. If they foster a healthier media landscape, including by reinvigorating local journalism, and regulate political parties, it will be much harder for populists to succeed.

This article appears in the Winter 2024 print issue of FP. Read more from the issue.

This article appears in the Winter 2024 issue of Foreign Policy. Subscribe now to support our journalism.