- amplification is usually thought of as the capacity for algorithmic systems to increase the reach of certain content beyond what it would have otherwise had. *for example*:
- allowing any spread of overtly harmful content,
- amplifying specifically inflammatory or divisive content, or
- disproportionately amplifying some type of user or perspective.
- conversely, repression might be thought of as *limiting* the reach of some content or accounts.
- this is hard to measure because there isn’t a good “baseline” for comparison.
- that is, how do we know how well this content would have performed *without* algorithmic amplification?
- [[lum, kristian and lazovich, tomo - 2023 - the myth of the algorithm]] suggests using, instead:
- “algorithmic exposure” → whether overtly harmful content was exposed or promoted at all.
- “algorithmic inequality” → comparing groups to each other to see whether one demographic or perspective is being amplified more than the other.
- amplification is distinct from [[what we think of as algorithmic radicalization is just how most of us surf the web|algorithmic radicalization]].
- radicalization is related to users being exposed to more extreme content over time.
- amplification is related to any type of content or profile being exposed to more users over time.
- radicalization is the depth and specificity of content; amplification is the reach of content.