Who Gets Seen Online? The Civic Stakes of Algorithmic Attention

A person can spend months developing a useful idea and still struggle to reach anyone. That familiar creator problem is now part of a larger public question: who gets seen when platforms decide what millions of people see next? Recommendation systems help match audiences with material they may value, but they also influence the news, arguments and communities that enter everyday life.

This is not an argument that algorithms are secretly choosing every belief. People retain agency, and creators make real choices about what to publish. But the systems that rank, recommend and remove material are not neutral windows. They are infrastructure. As social platforms become important places for civic information, their design has consequences for public discussion.

From finding a niche to finding a public

The original appeal of online publishing was relatively simple: a person could identify a subject, make something useful and find a community interested in it. Recommendation systems expanded that possibility. A small account can sometimes reach people far beyond its existing followers, while a viewer can discover a specialist who would be difficult to find through a traditional directory.

Ofcom, the United Kingdom’s communications regulator, describes this benefit as helping creators find audiences and reducing the time and cost of matching people with content. Its 2023 research also makes a crucial point: design choices can allow illegal or harmful material to spread when moderation systems fail to detect it. Ofcom recommends regular evaluations rather than one-off checks, using methods such as experiments, debugging exercises and user surveys. [3]

The same design that helps a thoughtful explainer travel can help a misleading or abusive post travel. That is a trade-off, not proof that all popular content is harmful. Popularity can reflect genuine usefulness, humor or emotional connection. The civic problem arises when the incentives for attention reward speed, outrage or certainty more reliably than accuracy and context.

A new layer of news distribution

The stakes are clearest when creators discuss public affairs. Pew Research Center’s study, published November 18, 2024, surveyed 10,658 U.S. adults and analyzed 500 popular news influencers across five major platforms. It found that 21% of adults regularly get news from social-media news influencers, rising to 37% among adults ages 18 to 29. Among those who use influencers for news, 65% said the creators helped them better understand current events. [2]

Those figures do not show that influencers are replacing journalism, or that their audiences are uninformed. They show that civic information is increasingly encountered through personalities and formats that feel direct, conversational and familiar. Most of the influencers in Pew’s sample had no background or affiliation with a news organization. That can bring valuable lived experience and a wider range of voices. It can also mean that standards for sourcing, corrections and separation of reporting from opinion vary widely.

Pew found that people receive a mixture of basic facts, opinions, breaking news and humorous posts. That mixture is part of the appeal: a creator can explain a complicated issue in language that feels accessible. Yet it can blur categories. A confident interpretation may be remembered as a verified fact, especially when the platform presents it beside professionally reported work without much context.

A viewer’s concern is therefore practical, not abstract. If a feed is the place where someone first hears about a local election, a public-health dispute or a war, how can that person tell what is known, what is disputed and what is simply a creator’s view?

What transparency can and cannot solve

Governments are beginning to treat platform design as a matter of public accountability. The European Union’s Digital Services Act (DSA) gives users rights including explanations for content removals, appeal routes and, on very large platforms, an option for a non-personalized feed. The European Commission says platforms with more than 45 million monthly users in the EU must assess systemic risks, including threats to fundamental rights, media pluralism and electoral processes. [1]

These measures address an important imbalance: platforms know far more about their ranking systems than users, researchers and many affected communities do. Clear explanations and access to meaningful data can make it easier to challenge errors and study harms. A chronological or non-personalized option can also give people a way to compare an algorithmic feed with a less tailored one.

Transparency, however, is not the same as accountability. A long report about a system’s inputs may be technically complete but useless to an ordinary user. Publishing a list of content categories does not reveal how a recommendation changes across age, language or location. Nor does disclosure by itself establish whether a design is causing measurable harm.

UNESCO’s Guidelines for the Governance of Digital Platforms, developed through a multistakeholder process, frame the issue through both safety and human rights. They call for transparency, accountability, independent oversight, human-rights due diligence and user tools for critical engagement. They also warn that poorly designed regulation can weaken civic space or become censorship. [4] That warning matters: reducing harmful amplification should not become a pretext for suppressing unpopular but lawful speech.

The competing concerns

One concern is that stronger duties could burden smaller services, slow experimentation or encourage platforms to remove too much material. Moderation is difficult at scale, and a rule designed to limit disinformation may accidentally suppress satire, minority languages or legitimate political dissent. Platforms also serve different countries and cultures; a single definition of “harmful” can be imposed unevenly.

The opposing concern is that leaving design decisions almost entirely private treats a public-facing infrastructure as if it had no public effects. Users cannot meaningfully consent to systems they cannot understand, and creators may feel pressure to exaggerate simply to remain visible. When platforms change ranking rules without clear notice, a livelihood, a community and a source of civic information can all be disrupted at once.

These concerns are not symmetrical in every case. Freedom of expression does not require a platform to recommend every post to every person. At the same time, safety does not justify secret rules, arbitrary enforcement or government control of lawful viewpoints. A credible approach needs independent review, accessible appeals, privacy-protective research and evidence about actual outcomes rather than assumptions about technology.

What a healthier information environment could look like

For users, the most realistic goal is not to master an algorithm. It is to build a few points of verification into ordinary habits: check the date, look for a primary document, compare a claim with more than one credible source and notice whether a post is reporting, analysis or advertising. UNESCO emphasizes media and information literacy that empowers users and narrows the participation gap between people who can critically create and use media and those who cannot. [4]

A recognizable lens on manufactured attention

Renée DiResta is a researcher and public writer whose work examines influence operations, networked information systems, and the ways online crowds can create a sense of consensus. Her relevance here is not that every popular post is coordinated or that one researcher can decide what is true. It is that her public work gives readers a name for a familiar cultural feeling: a claim can appear unavoidable online because a platform’s incentives, a network’s activity, and a creator’s format are reinforcing one another.[5]

That lens makes the article’s central distinction more concrete. People can make genuine choices and still encounter a feed shaped by forces they cannot see. The responsible response is neither conspiracy thinking nor blind trust in engagement numbers. It is to ask what is being amplified, by whom, with what evidence, and whether a recommendation has been mistaken for a public consensus.[5]

For creators, credibility can be a competitive advantage. Showing sources, correcting mistakes visibly and labeling opinion does not guarantee reach, but it gives an audience a reason to return beyond the next emotional spike. For platforms, the baseline should be understandable explanations, meaningful user controls, regular risk testing and independent scrutiny that protects personal data.

The public conversation about recommendation systems should also include people who are often missing from design debates: local journalists, educators, disability advocates, young users, researchers and communities working in less widely served languages. Their experience can reveal harms that aggregate engagement numbers conceal.

The desire to find one’s audience is human and legitimate. The question is whether attention should be allocated only by systems optimized for engagement, or by systems subject to public rules that protect both participation and rights. There is no perfect feed. But people can reasonably ask to know why something is being shown, to challenge consequential mistakes and to choose a less personalized path.

References

[1] European Commission, “The Digital Services Act,” last updated October 1, 2026. https://digital-strategy.ec.europa.eu/en/policies/digital-services-act

[2] Pew Research Center, “America’s News Influencers,” November 18, 2024. https://www.pewresearch.org/journalism/2024/11/18/americas-news-influencers/

[3] Ofcom, “Evaluating recommender systems in relation to illegal and harmful content,” July 6, 2023. https://www.ofcom.org.uk/online-safety/safety-technology/evaluating-recommender-systems-in-relation-to-the-dissemination-of-illegal-and-harmful-content-in-the-uk

[4] UNESCO, “Guidelines for the Governance of Digital Platforms,” accessed October 2026. https://www.unesco.org/en/internet-trust/guidelines

[5] Renée DiResta, official profile and writing, https://www.reneediresta.com/about/; U.S. Senate testimony on online influence operations, August 1, 2018, https://www.intelligence.senate.gov/wp-content/uploads/2024/08/sites-default-files-documents-os-rdiresta-080118.pdf; public profile: https://www.instagram.com/renee.diresta/ (accessed October 9, 2026).