At a guess: The articles are stored in some oldass MySQL DB that doesn't support Unicode, and they've somehow never gotten around to migrating the tables to a new schema.
So it’s entirely possible that they are using a fully up-to-date database, but the software depends on using a database collation that doesn’t support Unicode. I’ve seen software like that in active use even now. It can definitely get frustrating, and it’s not that hard to fix if you can just get some management buyin.