I do not know if what you are asking is possible, but you could always change the encoding from UTF-8 to ANSI (for example, using Notepad) before uploading.
Yes I did that. And that works as a workaround. I just think Optimizely should fix this in their export logic, so I do not have to remember to use this workaround every time. (The first time I didn't, and then I have to clean up).
That file is exported from the Search & Navigation service itself. So it would have to be changed for all sites that use it.
I doubt they would do that, since it needs to support languages with different character sets with a single solution.
Does importing not work properly with UTF-8?
The workaround explained by Thomas Hensrud Gulla does work! So changing the format manually before importing works.
When exporting synonyms from /EPiServer/Find/#optimize can you change the character set the file is created with, from UTF-8 to ASCII. This will make think easier when I (I'm Norwegian) imports the synonyms because we have synonyms with æøå!