They're just scraping page titles and occasionally short excerpts; most people believe that's fair use.