How social media tilted the public square toward the extremes—and helped populists surge

For years, analysts have searched for a single, simple reason behind the rise of populist movements. Immigration, stagnant wages, culture wars, and distrust in institutions all play their part. But one driver is often misunderstood: the way social media has quietly dismantled the guardrails that once favored moderation in public life. Traditional outlets—however imperfect—were built around editorial standards, professional norms, and a culture of balance that gave the broad middle a loud voice. Social platforms inverted that logic. In feeds optimized for attention, extreme views travel farther and faster than cautious ones, and those who master the mechanics of outrage routinely outcompete those who plead for context.

This is not just a morality tale about “bad content.” It is a story about incentives. For more than a decade, the core metric of the modern internet has been engagement: clicks, comments, shares, watch time. That metric rewards novelty and intensity. In practice, outrage beats nuance, certainty beats doubt, and a searing one-liner outperforms a sober paragraph. The result is a public square that behaves like a popularity contest in which the loudest voices are structurally advantaged. Populist entrepreneurs—on the left and the right, though the right has been more unified in its critique of institutions—have understood this better than most politicians and much of the press.

The great unbundling of gatekeepers

Before social feeds, professional editors, producers, and public broadcasters acted as de facto referees, curating facts and toning down extremes. Their role was never neutral—it reflected commercial pressures and cultural blind spots—but the system made it difficult for fringe narratives to dominate the conversation. Social media “unbundled” that model. Anyone can publish; anyone can go viral; anyone can assemble an audience larger than a mid-sized newspaper. The loss of friction is a democratic triumph in one sense and a civic challenge in another.

Empirical research over the last few years adds contour to what many users feel in their bones. Studies of recommendation systems and user behavior show that partisan and sensational material tends to earn more engagement than moderate posts; once users start down a path, the platforms’ personalization engines learn to serve more of the same. Separate research finds that people are unusually attracted to politically extreme voices within their own camp—call it “acrophily,” the pull of the peaks—which further skews attention toward the fringes. Even when platforms do not explicitly nudge people toward intolerance, the ecology of incentives does. In short: if moderation is a whisper, extremity is a megaphone.

The new populist playbook

In this environment, populists have rewritten political communication. The message is not merely that elites are out of touch; it is that elites are conspiring against the people—and that the proof is visible in your feed. A viral clip of a TV slip-up stands in for a corrupt media; a sensational statistic with no context becomes a referendum on the state. When the response comes—from fact-checkers, regulators, or traditional outlets—it can be framed as censorship or collusion, completing the narrative arc. The platforms’ architecture does the rest, rewarding the emotionally evocative and the highly certain.

Influencers—some partisan, some opportunistic—have become the new intermediaries between politics and the public. For younger audiences especially, TikTok-style short video and personality-driven news explainers often beat formal journalism at reach and recall. The consequence is not that people hear only extreme ideas, but that moderation is under-represented in the places where attention is allocated. A handful of hyperactive accounts can set the agenda for millions; the vast majority of users observe, react, and occasionally amplify.

Guardrails rolled back, guardrails reimagined

Platform policies have been anything but static. During the pandemic and the election cycles that followed, the biggest companies stood up aggressive trust-and-safety rules, partnered with fact-checkers, and limited the distribution of content judged misleading or harmful. In 2025, some of those measures were rolled back or reimagined. Meta, for example, said it would take a more “personalized” approach to political content and move away from heavy-handed demotions. X (formerly Twitter), after staff and policy overhauls, has emphasized community-driven context over centralized moderation. Supporters call these shifts a win for free expression; critics see a permission structure for chaos.

Regulators, meanwhile, are trying to build external guardrails. Europe’s Digital Services Act (DSA) aims to force greater transparency over recommendation systems and quicker responses to illegal content. Investigations have focused on how platforms rank, recommend, and label information, and whether they adequately mitigate systemic risks. Across the Atlantic, Washington has sent mixed signals: wary of foreign mandates shaping American speech norms, U.S. officials have warned companies not to weaken encryption or over-censor to satisfy European rules. The result is a regulatory tug-of-war over who gets to set the safety rails for a global internet.

Why extremes gain ground online

The over-representation of extremes on social media is not a conspiracy so much as a byproduct of three dynamics:

First, engagement is asymmetric. Anger, fear, and moral certainty produce more immediate reactions than ambivalence or doubt. A post that flatters your identity or enrages your in-group rivals is more likely to be shared than a careful explainer with caveats. Even modest algorithmic tweaks can’t neutralize the basic fact that emotionally charged content is more contagious.

Second, status flows to the bold. In social spaces where attention is a currency, users learn quickly that sharper takes earn more followers. Politicians and pundits who might once have calibrated their tone to win over swing voters now shape messages to excite a base that can propel them into constant visibility. The feedback loop turns a style of expression into a strategic necessity.

Third, fragmentation erodes common frames. When everyone’s feed is different, the common reference points that once underpinned national debates shrink. The center becomes harder to see not because it doesn’t exist, but because it has a weaker distribution channel. Over time, the very idea of a shared reality looks like a partisan claim.

The costs—to parties, platforms, and publics

Mainstream parties have struggled to adapt. Many still communicate as though nightly news bulletins set the agenda. Populist challengers, more fluent in the grammar of the feed, dictate the tempo. Traditional media has also been pulled into the vortex: what starts on niche accounts often becomes a TV segment by day’s end, forcing legacy organizations to chase and debunk rather than to set the conversation.

Platforms face their own bind. The very features that made them addictive also made them civically combustible. Dial down the engagement engine and growth slows; dial it up and social costs mount. Attempts to build “softer” feeds—more friends-and-family posts, fewer politics—risk accusations of shadow bans or bias. Greater transparency about ranking and recommendation systems helps, but disclosure alone does not change the underlying incentives.

For citizens, the tax is confusion and exhaustion. Many people retreat to lighter content or exit the news altogether. The paradox is that even as trust in institutions continues to slide, dependence on the platforms that disrupt trust has grown—especially for those who primarily encounter current affairs via short video and influencer commentary. In this setting, political entrepreneurs who thrive on simplification gain a structural edge over coalition-builders who need time and nuance.

What would rebuilding the guardrails look like?

A return to a tightly gated public sphere is neither realistic nor desirable. The task is to make abundance governable without turning expression into a bureaucratic permission. That starts with design. Platforms could offer friction by default on highly viral political posts—slower sharing speeds, more context on disputed claims, and optional tools to diversify the viewpoints in a user’s feed. Auditable risk assessments of recommender systems, already contemplated by European rules, could become a global norm. And there is room for civic product thinking: giving users the ability to tune for evidence, not just excitement, would be a meaningful start.

News organizations also have agency. They can treat platforms not merely as distribution pipes but as places where journalistic values need to be translated for new formats and younger audiences. That means investing in creators who can tell complex stories with clarity as well as in verification teams that work at feed speed. Public broadcasters and local outlets, uniquely positioned to serve the middle, can collaborate on shareable explainer content without abandoning depth.

Finally, politics itself has to adapt. The lesson of the past few years is not that moderation lacks a constituency; it is that moderation lacks a delivery system. Parties that want to resist the pull of extremes need credible messengers online, not just disciplined press conferences offline. They also need rules of the game—within parties and legislatures—that reward cross-faction bargains. Guardrails, in other words, are not just code and law. They are habits, incentives, and norms that must be rebuilt in the places where people now live their political lives: on their phones, inside the feed.

The populist surge did not arrive ex nihilo, and it will not vanish with one regulatory action or algorithm change. But the same network effects that elevated the extremes can, with will and design, elevate better behavior over time. If the past decade was about discovering what unbounded speech looks like at platform scale, the next one is about learning how to keep the democratic virtues of openness while restoring the civic advantages of moderation. That is not censorship. It is maintenance—of the commons we all share.