Good catch! It's definitely encoding/locales causing this, but I'm not quite sure which component would be entirely responsible -- if it's a bug in our ubuntu-mate-welcome, the json module, or a system config issue.
UTF-8 would be the encoding the software would expect to be processing with. I believe you can find out like this:
echo $LANG
en_GB.UTF-8
Coincidentally, the other week I put Ubuntu in a container on my Android device (armhf / armv8l), and a Ruby application had problems reading certain files. Turns out I had to put this in ~/.bash_aliases to fix it:
export LC_ALL="C.UTF-8"
export LANG="en_GB.UTF-8"
export LANGUAGE="en_GB.UTF-8"
Perhaps you may need something similar to prevent this happening on that device?