7 Societal impact
Recommender systems and social media have moved far beyond being simple tools for communication; they now shape everyday life, influence emotions, and alter how people make decisions. The chapter shows that online activity does not stay online: it changes what people feel, what they buy, where they get information, and even how they think about politics and society. Because these platforms optimize for engagement, their effects often extend to non-users as well, creating broad social spillovers that are hard to avoid.
A key theme is that seemingly small online influences can have meaningful real-world consequences when scaled to millions of people. Experiments on emotional contagion and voter turnout suggest that the content people see can affect mood and civic behavior, even if the measured effects are individually small. The chapter also warns that these systems can amplify existing social tensions rather than create them from scratch, accelerating polarization, outrage, and misinformation while giving powerful companies enormous influence over public discourse.
The most serious examples show how engagement-driven systems can help harmful content spread, including in cases of ethnic violence and genocide, when divisive material is boosted and moderation is weak or understaffed. At the same time, the chapter acknowledges that online platforms can also help people organize for collective action, as seen in major social movements and hashtag campaigns that spread awareness and coordination. Overall, the chapter argues that recommender systems reshape culture, news, and creativity by rewarding what is attention-grabbing, often at the expense of nuance, shared reality, and long-term social well-being.
Summary
- Studying the impact of recommender systems and social media in the abstract is hard, as their diffusion coincided with the rise in the use of such technologies.
- The Facebook studies on emotional contagion and voter boosting give insights into the role of recommender systems.
- By promoting emotionally charged content, platforms surface more extreme content—including harmful ones—over time.
- In extreme cases, this can lead to genocide. In Myanmar, Facebook ignored concerns of local activists and let misinformation and harmful content spread unchecked.
- Social media offered a new way for people to organize online. Recommender systems expanded the reach by showing content to new audiences.
- The underlying technology is agnostic of the specific content, so it can equally spread movement who advocate for positive changes (such as #BlackLivesMatter, and #MeToo), and harassment campaigns or full-blown conspiracy theories.
- As social media became our door to the internet, and recommender systems the default on feeds, they have a significant effect on how we form opinions on the world.
- Personalization can lead to a lack of shared reality and an informed public, needed to discuss the pressing issues.
- Content creators adapt the format of their content to maximise their chances of being promoted by algorithms, giving a sense of cultural flattening and content homogenization.
FAQ
What is meant by the societal impact of recommender systems?
It refers to the changes in society that can be attributed to recommender systems and related internet technologies, including effects on emotions, behavior, politics, culture, and information flows.
How can recommender systems influence our emotions?
Studies on social contagion show that the kinds of posts people see in their feeds can affect the emotions and tone of what they post themselves, though the measured effects may be small.
What is social contagion in the context of social media?
Social contagion is the spread of behaviors, ideas, or emotions through a social network, such as friends influencing one another online.
Can social media content affect political participation like voting?
Yes. A large Facebook experiment found that showing users a voting reminder, especially with photos and names of friends who voted, increased political mobilization and likely raised voter turnout.
Why is the influence of social media on voting considered important if the effect is small?
Even small percentage changes can matter at scale, especially in close elections where a relatively small number of votes can affect outcomes.
How did engagement-optimized recommender systems contribute to harmful content in Myanmar?
They helped amplify divisive anti-Rohingya content, which spread hate more widely in an environment with weak moderation, contributing to real-world violence and genocide.
Why are online platforms said to have no opt-out effect on society?
Because once enough people use them, their effects spill over to non-users too—for example, by changing news, culture, local businesses, or public discourse that affects everyone.
How do social media platforms help people organize online?
They provide tools for decentralized coordination, allowing people to share information, form movements, and organize protests without needing a central bureaucracy.
How have recommender systems changed news consumption?
They have shifted many people toward personalized feeds that often favor attention-grabbing or extreme content, which can weaken shared reality and reduce exposure to balanced journalism.
What is the effect of recommender systems on culture and creators?
They influence what people consume, which shapes taste over time, and they pressure creators to adapt their content to what algorithms reward, often leading to homogenized, viral-style content.
Hidden Influences ebook for free
