Use the P93081 "work with flat file encoding" app to set the desired codepage for the output of the CSV for the report and version in question. CP1256 is Windows-Arabic. Or set it to UTF8, but Excel sometimes has issues with UTF8 on csv import.
ah, I see you signature line -- 8.96 might be before they made UBE CSV export use that table, that is quite old.
Turn on UBE Debug log at level 6, see if it is fetching an encoding from F93081. Make sure is is the same datasource/owner that the client is setting -- if server map points to a different table than client, you may not be finding the record.
Also, if I recall correctly, your Windows user profile needs to be Arabic if you want Excel to recognize it as Arabic, independent of EOne.
Try loading the CSV in Notepad, then select all, then pick Format->Font and pick Arial Or Times (or any font with Arabic script encoding support), and in the Script drop-down, pick "Arabic" and see if it renders. That could indicate that the CSV is ok, it is Excel that is freaking out.
For Excel; Open Excel with a blank workbook. Click "Open" and then pick the CSV filename. If it just opens and doesn't ask you text import wizard questions, rename the file to .txt and try again. When you click "Open" and pick the file, in the Text Import Wizard, select File Origin dropdown setting of "1256: Arabic (Windows)" and see if it imports that way.