According to Harvard University Libraries, Google will scan not only books out of copyright, but also books still in copyright. It will not make those latter books available publicly--presumably until they have passed out of copyright.
Will this include books still in copyright?
Google will be scanning books that are in as well as out of copyright from the Harvard collections. Harvard-owned books in the public domain will be available in the search results. Google may choose to display descriptive catalog information for books that are still under copyright. We believe that Google's treatment of in-copyright works is consistent with copyright law.
I think this should be right, and I commend the Google/Harvard lawyers for proceeding on this basis.