I have looked into the US rules and come to the conclusion that as long as the
use is clearly and only "for scholarship" i.e. for research, including making
results available, and as long as it does not somehow compete with the source,
then the "fair use" doctrine protects the user.
Corpora can be configured so that they do not compete in any way with their
For added reassurance, a corpus access website can be set up so that anyone who
gets to the material has passed through a page that contains an agreement to use
the material only for research, and clicked on "I accept these conditions" or
something like that.
I don't think we should let these worries stop us from sharing. We can respect
the sources' interests without being silent.