Scans of public-domain work remain in the public domain under U.S. law, at least as currently interpreted. There's no Supreme Court decision on the subject, but the leading case widely followed is Bridgeman Art Library v. Corel (1999) (http://en.wikipedia.org/wiki/Bridgeman_Art_Library_v._Corel_...)
Access to network resources is another matter--- Bing might be violating Google's access policies if they slurped Google Books, especially if they evaded rate limits or ignored robots.txt. But they wouldn't be in violation of any copyrights. This makes it a bit of a cat-out-of-the-bag situation. If I download a copy of an 1878 journal from JSTOR, and email it to a friend, I'm violating JSTOR's terms of service. But if my friend turns around and posts the PDF online, he is not breaking any laws.
A similar situation holds with U.S. government classified documents. Whoever leaked the Pentagon Papers broke the law, and could be prosecuted for the leak. But once they were out, as public-domain government documents, it was not illegal for third parties to republish them.
Access to network resources is another matter--- Bing might be violating Google's access policies if they slurped Google Books, especially if they evaded rate limits or ignored robots.txt. But they wouldn't be in violation of any copyrights. This makes it a bit of a cat-out-of-the-bag situation. If I download a copy of an 1878 journal from JSTOR, and email it to a friend, I'm violating JSTOR's terms of service. But if my friend turns around and posts the PDF online, he is not breaking any laws.
A similar situation holds with U.S. government classified documents. Whoever leaked the Pentagon Papers broke the law, and could be prosecuted for the leak. But once they were out, as public-domain government documents, it was not illegal for third parties to republish them.