could perl ever use utf-16 internally?

I'm been thinking about this hypothetical question lately: could perl
ever switch its 'other' internal representation of strings from utf-8
to utf-16?

My initial guess is that it probably couldn't happen without breaking
a lot of apps. I think there are too many (sloppy) coding practices
which rely on the convenient fact that for characters in the US-ASCII
range, Perl will always represent those characters internally as
single octets.

Anyone have any thoughts on this?