This change would make the lives of your fellow Europeans a lot easier. Especially slavic languages heavily rely on accents.
The current solution to avoid possible filter bypasses is to simply remove all non-standard (a-z) characters from the message, but removing letters from every other word makes it quite difficult to talk.
My proposal would be to whitelist certain characters like á, à, â, ä and so on, but have them go through the filter as standard "a" character to avoid any bypasses; if that is possible on a technical level, of course.
It's no secret that the filter is in need of an overhaul anyway, so this might be a fix worth considering.