07 · Echo Chambers and Algorithmic Bias
The feed learns what keeps you scrolling, not what's true
Social media algorithms are designed to maximize engagement, which can trap users in echo
chambers by prioritizing content that aligns with their pre-existing beliefs — worsening
political polarization not through any single biased story, but through the cumulative shape of
everything a person sees over time.
How it works
Recommendation systems are built to maximize time spent and interactions, not accuracy or
balance. Content that confirms what someone already believes, or that provokes a strong
emotional reaction, tends to get more clicks, shares, and comments — so the algorithm learns
to serve more of it. Over time, a feed can narrow into a version of the world that consistently
agrees with the viewer, without any single post being false or any deliberate decision being
made by the person seeing it.
Why it works on us
People already have a natural preference for belief-confirming information — a well-documented
tendency called confirmation bias. Algorithmic curation doesn't invent this tendency, it
amplifies it at a scale and speed no individual editor ever could, quietly narrowing the range
of what a person encounters until the narrowing itself becomes invisible from the inside.
Signs to watch for
- Rarely encountering a view you disagree with in your main feed, even on contested topics.
- Recommended content on a topic gradually escalating in intensity or certainty over time.
- The same claim or framing appearing from many different accounts within a short window.
- Feeling genuinely surprised by how a story is covered outside your usual feeds or apps.
- A sense that "everyone agrees" on something that, outside your feed, is actually contested.
Illustrative example
Two people searching the same neutral topic on the same platform see meaningfully different
results and recommended follow-ups, shaped by their own past engagement rather than by the
topic itself.