I'm a little late to the party, but here's a thread on teaching demonstrations. The discussion has already been going on in comments starting here, and Michael Cholbi tackles the topic at In Socrates' Wake here.
I've never done a teaching demo, but I've witnessed a few. My sense is that sending a reading may not hurt you, but conducting your demo in a manner that presupposes that they have done the reading will definitely hurt you, and not sending a reading will definitely not hurt you. So I think the safe play is not to send one. It's okay to base your demo around a particular reading; just don't base your demo on the premise that the students have read it.
It also seems to me that it would be a profound mistake to devote 12 minutes of your demo to a discussion in small-groups. What would such a discussion demonstrate? That you can effectively break students into small groups and let them discuss the material on your own? It will demonstrate that you were not prepared to give a teaching demonstration. Am I wrong?
The best teaching demo I have witnessed was on a historical topic, which was to be the candidate's primary upper-division teaching responsibility. The candidate had a particular issue to discuss, set out the issue in a clear way, explained how the relevant historical figures solved the problem, and explained the difficulties associated with these proposed solutions. He solicited student participation (one of the main difficulties was proposed by a student), made us feel comfortable, and was funny and engaging. In the final section of the demo, he connected the historical topic to a closely-related current debate. This candidate got the job.