The BBC has just mentioned that "existing entries in non-Latin text are sometime being duplicated, as users will search in Latin text and not see a result in Cyrillic or Japanese".

We are showing the primary English alias for the users in inline searches, but not in other searches on the site. Since the BBC is usually searching for the entity so they can enter the MBID into their own software, inline search tools won't help them. We should probably display the primary aliases in the standard search results too, so that someone just searching normally isn't given a list of Cyrillic-only results (this is half-"solved" for artists by showing the sortname, but that's mostly a hack, and we should probably stop showing it once we display the alias).