To the extent that it automatically gets populated, it contains copyright or licensing information for the resource. It's not used for anything (e.g. won't appear in citation styles), though, so it's purely for the user's own records.
It just depends on the site you're importing from -- in some cases Zotero grabs that information. I think e.g. when importing from the Wikimedia Commons, but also in the case of some commercial publishers. Looking at my library, I'm seeing e.g. Wiley, Springer, Nature, Washington Post (copyright statements) as well as github and a couple of blogs with good metadata (license)
It's not used for anything (e.g. won't appear in citation styles), though, so it's purely for the user's own records.