The attention economy is incompatible with democracy

What's this about?

People disagree about whether the attention economy, where apps fight for your time, can fit with democracy (rule by voters).

The key worry is that apps may push loud, angry posts instead of useful facts.

What supporters say

• Posts with strong moral words and big feelings often spread farther than calm, careful posts.
• Anger and fear can make people share more, so apps may reward posts that upset people.
• False news spreads faster and wider than true news on Twitter, mainly for public issues.
• Online feeds can sort people into groups and make them dislike people with other views.

What critics say

• Studies do not prove that angry posts directly change votes or end democracy.
• Fake news did not clearly have a large effect on votes in the 2016 US vote.
• One Twitter test found more group splits for some users, but it may not fit all apps.
• People may choose posts that match their views, so apps may not cause all the harm.

The bottom line

The evidence shows a real, ongoing clash between attention-based apps and healthy public debate.

But we cannot say these apps will always destroy democracy in every place or with every app design.

The claim is that an online economy built to capture attention cannot live alongside democracy. The evidence points to a serious and recurring conflict, but not to an inevitable collapse of democratic politics in every country or under every platform design.

The case for

The strongest argument is that attention-driven platforms reward exactly the kind of political material that can damage careful public debate. Research finds that posts using moral and emotional language spread more widely, while highly arousing feelings such as anger and anxiety are linked to more sharing. That creates a built-in incentive for content designed to provoke rather than inform. Outrage travels well online, even if the studies do not prove that it directly changes votes or causes democracy to fail. 1

The information problem is especially stark. False news has been shown to spread farther, faster, deeper and more broadly on Twitter than true news, particularly when the subject is politics (see Figure 1). 2 This suggests that novelty and emotional force can beat accuracy in a system where shares, clicks and viewing time are valuable. Evidence from the 2016 US election did not establish that fake news had a large effect on voting, but it remains a clear weakness in the political information environment.

There is also reason to doubt that high engagement leads people toward better understanding of their opponents. In one randomized Twitter experiment, seeing opposing political views increased polarization among some users, especially Republicans. Broader research reviews have linked social media to selective exposure, ideological sorting, hostility and polarization. 3 Still, these findings have limits: the Twitter result may not apply across all platforms or populations, and observational studies cannot always separate a platform’s effects from users’ pre-existing beliefs.

At a wider institutional level, official and civil-society assessments warn that recommendation systems and targeted distribution can be exploited. They cite coordinated political manipulation, censorship, election threats, weaker information quality and falling trust. Platforms optimized for attention can give a small number of commercial companies major power over what political issues people see and discuss, raising concerns about equal political voice and public accountability. 45

The case against

The main objection is that the evidence does not show that attention-based platforms make democratic breakdown unavoidable. During the 2020 US election period, coordinated Facebook experiments made major changes to what users saw and how they engaged with content. Yet those changes did not produce large short-term shifts in political attitudes or most measures of polarization. 6

That result fits a broader picture in which media effects are mixed and depend on many forces beyond platform algorithms. Users choose what they consume, while political parties, traditional media, social networks and existing beliefs also shape outcomes. A change in online exposure or engagement is therefore not the same thing as a lasting change in knowledge, tolerance, accountability or voting behavior.

Social media can also serve democratic purposes. It can expose people to political information, give more people a chance to speak, help groups organize, encourage discussion and support participation. 7 The evidence for these benefits is often based on correlation rather than firm proof of cause and effect, and the evidence on mobilization is less settled than some major research reviews. But the benefits are real enough to complicate any claim of outright incompatibility.

Nor is the current platform model fixed. Experiments repeatedly find that brief prompts asking users to think about accuracy can improve the quality of news they say they intend to share. Policy proposals also point to transparency, stronger journalism, more diverse media ownership and better governance as ways to reduce information risks. 8 Such interventions may not work equally well in every setting, but they show that design choices and rules can redirect some attention-driven behavior toward accuracy.

The bottom line

The attention economy is strongly in tension with democracy, especially with the need for accurate information, equal voice and thoughtful public debate. Its incentives repeatedly favor emotional content, allow falsehoods to spread effectively and create openings for manipulation.

But the evidence does not prove that it is inherently incompatible with democracy in every form or setting. Effects vary by platform, country and political institutions. The risks are likely greatest where trust is low, journalism is weak, manipulation is organized and governance fails. The biggest unanswered question is whether these harms build into lasting, system-wide democratic damage over years and across countries—even when platforms are redesigned and institutions respond.