I have a google alert for one of my open source projects (APSW) and every day almost they email me one or two sites where it has popped up. Virtually every one

I have a google alert for one of my open source projects (APSW) and every day almost they email me one or two sites where it has popped up. Virtually every one has a bizarre name, and on going to the site it is content obviously generated using Markov chains from other sites and incomprehensible to humans. Of course there is no way to tell Google that these aren't real sites (I tried) and distinguishing real text from Markov chains using a computer is hard. I keep wondering what it will take for them to figure out bad content, both for the computer (javascript etc) and for people (english etc).