SharePoint Search: full-text queries matching fragments of UniqueId / ListId document properties

We found out the hard way that, by default (as this behavior is reproducible on SP2019 vanilla farm installations), SharePoint Search includes a tokenized version of documents' UniqueId/ListId into its full-text index. A screenshot will explain the situation better than a thousand words:
We found no way to configure this behavior, as UniqueId/NormUniqueId/ListId/NormListId managed properties aren't searchable, and ows_UniqueId and ows_ListId crawled properties aren't included in the full-text index, so "Search Schema"-wise they shouldn't already be there.
Is there some way to exclude those properties from the index? We have some custom metadata that must be searchable and is similar to the guid parts being matched, so this is causing a lot of irrelevant results to be returned to users.