Two people can follow the same accounts, log in at the same time, and see different feeds. Personalization adapts what each person sees to their preferences; past interactions with a platform and its content can influence the result. section 7.5
Recommender systems use personalization to help people find appealing content among many options. Platforms can also optimize feeds for business objectives such as time spent or engagement. What keeps someone engaged, however, is not necessarily what they find valuable. section 7.5 section 1.1.1 section 7.5.2
A movie's many faces
Personalization can affect presentation as well as selection. A streaming service might promote the same movie with an action shot to one viewer and a romantic image to another. This raises a question: when does tailoring artwork make the movie misleading? Stacia Brown asked whether Black users were being shown artwork featuring Black actors with minor roles; Netflix denied tailoring artwork to a specific race. section 6.1
In news, a feed that offers only what a person wants to see may crowd out important but less immediately appealing stories. People may also encounter different accounts of events, making a shared understanding of public issues harder to sustain. A range of perspectives is valuable, but it is not the same as choosing content solely to maximize engagement. section 7.5.1 section 1.1.1
Personalized feeds are also difficult to audit: researchers cannot inspect a single public feed to see what everyone receives. At the same time, creators may adapt their content to platform incentives, contributing to short-lived trends and more homogeneous content. Personalization therefore involves choices about user value, platform goals, representation, and public accountability. section 7.5.2 section 7.6










