I mean from the www into steemit of course.
Also unless there is google like power behind the cheetah bot there is absolutely no way to make sure that people do not copy. I suppose it tries a google search for some sentences and tries to find copied content that way, but what if the person trying to copy get's his information before it get's listen on google? What if he copies it from a source that isn't trackable to some online book or website url.

It's unavoidable that content will be brought in from the outer www.
straight up copied content