ArsTechnica seemed to sum it up best:
Today, it is expected that Google will announce an agreement to scan and create databases of works from five major libraries. According to news reports, Google will digitize all volumes in the University of Michigan and Stanford University library systems along with parts of research libraries at Harvard, the New York Public Library, and Oxford University in England. More information on the scope of projects at the individual institutions can be found at news.com. The project looks to be an extension of Google Print and Google Scholar, while reaching all the way back to the Stanford library digitization project where Google originated.
TeleRead, always an advocate for digital libraries, was right on top of it.
Remember the C word from the dotcom boom–the convergence of different media? Well, how about the convergence of books and the Net, in ways going far beyond Gutenberg and similar efforts?
But TeleRead is always conscious of copyright and other limitations on use and access, and that’s right where CopyFight went with the story:
Having this material available has the potential to place in the foreground the importance of the public domain. Further, Google may have the resources and incentive to figure out comprehensively which post-1923 books are already in the public domain for failure of formalities.
But that was the rosy quote. Most of the concern is about how Google can escape accusations of piracy if they post even portions of copyrighted books. The search engine giant is already facing a lawsuit over links to copyrighted images.