Now open your phone and take a look at whatever comes up. A video that someone suggested. A piece of news. A post from a person you hardly know yet follow. Take a moment to imagine that none of those decisions were truly yours. Something chose what to show you in what order, what to reveal and conceal, and what to maintain your attention for a short while. It’s an algorithm that has been influencing your daily media experience for years, mostly without your awareness or permission.
Although content curation algorithms are not new, their size and complexity have advanced to the point where they operate more like customized media editors than straightforward filters. These editors work incredibly quickly, optimizing for a specific objective that is rarely the same as yours. When you launch a social media app, your objectives could be to pass the time, stay informed, or catch up with acquaintances. The algorithm wants to keep you interested for as long as it can. There are moments when the two goals overlap. They differ in significant ways.
For many years, the fundamental mechanics have been quite well known. What you click, how long you pause a movie, what you share, and what you skip are all monitored by the algorithm. It puts you in groups with other users who have similar habits and feeds them comparable stuff. Because reactions indicate engagement, which is what the platform offers to advertisers, it gives priority to items that elicit significant reactions, such as comments, shares, or statements of outrage. A video that makes people laugh, a narrative about a local school board meeting, and a viral clip that makes people feel something they can’t quite put their finger on all directly compete with one another. Regardless of whether something is true or helpful, the algorithm tends to favor things that elicit strong emotions in users.
Ten years of automated curation have been studied, and the results are unsettling but not very surprising. According to a 2024 review published in a European communications journal, algorithms more consistently reinforce preexisting beliefs by creating what researchers refer to as “filter bubbles,” which are environments where the content you see increasingly aligns with your preconceived notions. However, they can occasionally expose users to viewpoints they wouldn’t have sought out themselves. Because emotional content drives the engagement metrics the algorithm is optimizing for, it has been demonstrated in numerous audits that Twitter’s algorithm amplifies emotionally charged, low-credibility content more consistently than calm, factual stuff.
Facebook discovered in 2016 that its users were publishing more professionally generated news and less original personal content. This was an unanticipated result of curating, which the company later attempted to rectify with an algorithm upgrade in 2018. The change, which was meant to highlight “meaningful social interactions,” ended up boosting content that caused division in various ways, raising issues that the firm has been dealing with ever since.
It is especially important to consider the consequences on younger users. A 2025 longitudinal study found that two-thirds of Australians between the ages of 12 and 16 now get their news mostly from social media. Instead of using a system that is intended to provide people with factual information, they are obtaining their view of the world through one that is specifically made to maximize the amount of time they spend on a platform. While most teens are consuming algorithmically tailored content without really understanding why they’re seeing what they’re seeing, teens who utilized multiple platforms demonstrated a stronger awareness that algorithms were altering their feeds, suggesting that exposure itself taught some skepticism.
In some places of the world, a regulatory response is gradually emerging. Large platforms must provide users with an algorithm-free, chronological feed alternative in accordance with the EU’s Digital Services Act. That’s a significant intervention, but it’s unclear how many users will intentionally pick a less curated experience. Despite its flaws, the curation is typically more engaging than a raw chronological feed since it’s designed that way. There is no such federal mandate in the United States. On this issue, the platforms have mostly maintained their autonomy, modifying their algorithms in response to public criticism while maintaining the fundamental engagement-optimization rationale.
