Related information has been used on written content since it's inception, and now Confluence is adding functionality for this as well. It is a feature that not always is welcome, however, but don't worry, you get plenty of control to decide if it should be used or not. Not just per space, but even down to page level.
While we don't have specifics on the algorithms that make this new Related information work, but we can probably assume that labels and keywords in title and headers in the content is used somehow. We do know that the related pages will honor permissions, so you can only see pages you have access to. The official information have this to say:
Related pages are automatically generated suggestions for further reading, based on pages that are viewed or are organized together. This helps you gain context on collaborative work faster, and quickly discover relevant articles in a knowledge base, even when the pages aren't linked or organized together.
When this rolls out, it will be turned on by default, which may be a bit annoying. Fortunately, you can turn this off for whole spaces in the Space settings under Space settings > Related pages. You can also turn this on or off on individual pages under More actions (•••), then Advanced details > and either Show related pages or Hide related pages.
In the first iteration of related pages, it will not support Blogs or Jira Service Management customer portals. While this might seem strange as these two areas are where you normally would use this feature, but I think it is a good thing as they can tweak the algorithms a bit before taking that step. This should help make that feature better when it is released. This should help a lot with support if the algorithm is good enough, which I think it will be once they get some data from Confluence usage and tweak a bit first.