Aw, come on. Abstracts are not generated automagically by snipping a headline and the opening text of the paper or book.
They are either provided by the author, or written by a real live person that took the time to read the work in question.
Again, I would ask the same question I asked previously (which remains unanswered). Would you be OK with me scraping your site? I promise, I will just grab the <h1> and opening <p>.
Is that fair use?