fbpx

Human Progress Logo Human Progress

01 / 05
We’re Living in a Split-Screen America

Blog Post | Communications

We’re Living in a Split-Screen America

The evolution from broadcast news to personalized feeds has fractured how we see the world, but progress is possible.

Dylan Tweed —

Summary: Americans once shared a common media landscape, but the rise of personalized digital feeds has splintered that reality into partisan echo chambers. Social platforms now amplify outrage, reinforce tribal instincts, and erode agreement on basic facts. While there is no easy fix, reforms in design, digital literacy, and cultural norms offer hope for a more truthful and united public discourse.

“And that’s the way it is.” At least, that’s the way it was. When Walter Cronkite closed his nightly broadcasts with those words, America was a foreign country. At the height of broadcast news, Americans had differences of opinion but agreed on a basic set of facts about what was going on in the country and the world. Anchors like Cronkite, voted in 1972 by Democrats and Republicans alike as the most trusted man in America, aimed to be impartial and to win bipartisan credibility. But as partisan cable news and talk radio came to prominence in the 1990s, basic agreement on the facts began to erode. And with the rise of social media, it splintered entirely.

Platforms like Facebook, YouTube, TikTok, and Twitter personalize content to maximize engagement (time spent on an app, posts liked and shared), showing you what you want to see. That reinforces users’ existing beliefs and limits exposure to opposing views. Strikingly, a Meta-commissioned study of 208 million users during the 2020 U.S. election cycle showed that liberals and conservatives on Facebook encountered almost entirely non-overlapping news sources. Once a social media user spends time looking at political content on one of these platforms, he or she is fed more and more of the same. Far from the broadcasts of the mid-century, modern news is delivered via increasingly bespoke “narrowcast.”

This political siloing is not trivial. Americans now inhabit split-screen realities. In one 2023 Gallup poll, 90 percent of Republicans believed crime was rising, while 60 percent of Democrats believed it was falling. On climate change, a 2021 survey showed a 56-point partisan gap in beliefs about whether humans have a serious impact on the climate system (compared to a 16-point gap in 2001). In 2024, 44 percent of Democrats rated the national economy as “excellent or good,” compared to only 13 percent of Republicans, despite the same underlying economic conditions. The gap wasn’t driven by personal finances, but by partisan interpretations of identical economic indicators. These are not differences of opinion; they are incommensurable beliefs about the state of the world.

But platforms don’t just feed us headlines that align with our politics. They also bait our strongest emotions. In 2017, Facebook began weighting “angry” reactions five times more heavily than “likes” when floating posts to the top of our feeds. That same year, a study found that each additional moral-emotional word in a tweet (think “shameful,” “detestable”, “evil”) significantly increased the likelihood of it being shared and reshared.

This platform design calls up ancient instincts. Humans evolved to detect threats to the coalition, to signal our group loyalty, and to rally allies against rivals. A tweet calling someone “abhorrent” isn’t just an opinion; it’s a tribal call to action. And because these platforms so reliably elicit our ire and impel us to spread it to others, they’ve become outrage engines.

They create sealed chambers that echo our anger, where contrary evidence is unlikely to penetrate. Carl Sagan now sounds prescient when he warned in 1995 of a future where Americans, embedded in an information economy, would become “unable to distinguish between what feels good and what’s true,” leaving society vulnerable to illusion and manipulation.

And the consequences of the outrage engines don’t stop at our borders. In 2016, Russian operatives used fake personas on Facebook and Twitter to spread inflammatory memes targeting both liberals and conservatives. They didn’t need to hack anything. They simply exploited an information ecosystem already optimized for spreading partisan outrage.

What can be done? There is no single fix, but meaningful improvements are possible.

In a randomized study, older adults who received just one hour of digital literacy training from MediaWise improved their ability to tell false headlines from real ones by 21 percentage points. When Twitter added a prompt asking users if they wanted to read an article before retweeting it, people were 40 percent more likely to click through to the article before sharing it impulsively.

Choice helps too. In one study, switching users from a feed that had been personalized by the algorithm to one that showed posts in chronological order measurably increased their exposure to content across the political aisle. While it may not be a silver bullet, giving users the ability to choose their feed structure, including which algorithm to use, allows for opportunities to be exposed to contrary opinions and to peer outside the echo chamber.

But deeper change is cultural. A compelling case has been made that human reasoning evolved not to uncover objective truth, but to persuade others, to justify our own ideas, and to win arguments. That is why the habits of sound reasoning must be cultivated through norms that prize truth over tribal loyalty, deliberation over impulsivity, and the ability to make the best case for opposing views in order to oppose them on their merits.

This isn’t a call for censorship or government control of the news, nor is it a plea to go back to three-network broadcasting. The democratization of media has brought real benefits, including broader participation in public discourse and greater scrutiny of powerful institutions. But it has also made public life more combustible and has manufactured disagreements about factual questions. In a competition for attention, platforms are designed to maximize time spent on them. That means elevating content that provokes strong emotional responses, especially outrage, and targeting it toward the users most likely to react. The more incendiary the content, the more likely it is to hold us captivated.

What we are witnessing is not a failure of the market, but a particularly efficient version of it, albeit one that optimizes for attention, not accuracy. Personalized feeds, algorithmic curation, and viral content are giving people more of what they want. And yet, many Americans say they are dissatisfied with the result. In a 2023 Pew survey, 86 percent of U.S. adults said they believe Democrats and Republicans are more focused on fighting each other than solving real problems, and respondents across party lines cited political polarization as the biggest problem with the political system.

While online outrage bubbles may not qualify as a market failure in the technical sense, they are clearly a civic problem worth confronting. An information ecosystem optimized for attention rather than accuracy will reliably amplify division and distrust, even while giving users more of what they like to see and share. The incentives are working as designed, but the outcome is a fragmented public unable to agree on the real state of the world. If democracy depends on a shared understanding of basic facts of the matter, then reckoning with these tradeoffs is well worth our much-demanded attention.

Our World in Data | Communications

Number of People Without Electricity Has Halved Since 2000

“Most people in the world would think very little before flicking on the lights, charging a mobile phone or turning on a laptop to read this.

But that’s a very different reality from the almost 700 million people in the world who have no access to electricity. While this number is large, it has halved this century, falling from 1.35 billion to 675 million. You can see this in the chart.

However, this progress has been far from even. The number has fallen across all regions except Sub-Saharan Africa, where it has increased.

That doesn’t mean no progress has been made: the share of people in Sub-Saharan Africa with electricity has doubled, rising from 26% to 53%. But population growth has outpaced this expansion, meaning the number of people without electricity has still risen.”

From Our World in Data.

Harvard Gazette | Mental Health

Teen, Young Adult Suicides Fall After Crisis Hotline Shifts to Three Digits

“Suicide deaths among young adults and youth declined after a federal agency simplified the phone number for a national crisis hotline and increased resources, a new study says…

Patel, a clinical fellow in surgery at Harvard Medical School and surgical resident at Brigham and Women’s Hospital, said that when researchers first examined figures for all age groups, the lifeline’s potential impact appeared to be slight.

But when they broke down the data, they saw a significant decline among those age 15 to 34 — encompassing the high-risk teenage years — that had been masked by results in other groups.

The researchers noted a decline from both observed suicide deaths in 2022 and from predictions based on a long-term upward trend. In 2010, about 11 suicides per 100,000 were reported in that age group. By 2022, that had risen to nearly 18 per 100,000. Three years after the 988 number went online, however, that had fallen to approximately 15 per 100,000, according to the study…

In addition to the nationwide figures, state-by-state data also shows an association with the establishment of the 988 number.

The 10 states with the largest increases in calls after its establishment — 146.2 percent more — also saw a larger decline in suicide deaths, about 18.2 percent. The 10 states with the lowest call volume increase — about 23.6 percent — saw a lower, 10.6 percent decline.”

From Harvard Gazette.

arXiv | Communications

LLMs Show Promising Social Media Fact-Checking Capabilities

“Large language models show promising capabilities for contextual fact-checking on social media: they can verify contested claims through deep research, synthesize evidence from multiple sources, and draft explanations at scale. However, prior work evaluates LLM fact-checking only in controlled settings using benchmarks or crowdworker judgments, leaving open how these systems perform in authentic platform environments. We present the first field evaluation of LLM-based fact-checking deployed on a live social media platform, testing performance directly through X Community Notes’ AI writer feature over a three-month period. Our LLM writer, a multi-step pipeline that handles multimodal content (text, images, and videos), conducts web and platform-native search, and writes contextual notes, was deployed to write 1,614 notes on 1,597 tweets and compared against 1,332 human-written notes on the same tweets using 108,169 ratings from 42,521 raters. Direct comparison of note-level platform outcomes is complicated by differences in submission timing and rating exposure between LLM and human notes; we therefore pursue two complementary strategies: a rating-level analysis modeling individual rater evaluations, and a note-level analysis that equalizes rater exposure across note types. Rating-level analysis shows that LLM notes receive more positive ratings than human notes across raters with different political viewpoints, suggesting the potential for LLM-written notes to achieve the cross-partisan consensus. Note-level analysis confirms this advantage: among raters who evaluated all notes on the same post, LLM notes achieve significantly higher helpfulness scores. Our findings demonstrate that LLMs can contribute high-quality, broadly helpful fact-checking at scale, while highlighting that real-world evaluation requires careful attention to platform dynamics absent from controlled settings.”

From arXiv.

Cato Institute | Communications

Starlink Connects Millions of People in Argentina 

“When Javier Milei became president of Argentina in December 2023, one his first measures as part of a package of wide-ranging deregulations was to open up the economy to satellite internet. (I wrote about that and his broader deregulatory push here.)

At a meeting I attended last month with a small group of economists, Argentina’s Minister of Deregulation, Federico Sturzenegger, presented the graph above. It shows how satellite internet use exploded once the government lifted its ban, which had, until then, benefited a politically powerful local internet provider.”

From Cato Institute.