Looks like the EditorJourney schema records HTML tags in the page_title field. For articles (pages in namespace 0), the result appears to only consist of articles with a title that should be in italics, and the page_title field becomes <i>[obfuscated page title string]</i>.
For pages in other namespaces, this appears to mainly affect user and user talk pages (namespaces 2 and 3). There, we find examples using both the font and span elements as well.
Currently this is unlikely to affect our data analysis, where the combination of namespace and title is used instead. In other words, this is definitely a "nice to have" on the prioritization scale.