This[1] explains it. Essentially, to support unicode probably, you have to do extra work in your own code. (How do you handle invalid unicode? Should regex character classes match the unicode versions, or just the ascii versions? etc.) Additionally, certain builtin functions don't already support unicode, and modifying them to do that would be a breaking change. For more info, see the py2->py3 unicode change, and how much of a mess that caused.

1. https://stackoverflow.com/questions/6162484#6163129