This, in my opinion, would be even easier for Amazon to detect than gibberish sentences pieced together. There are plenty of companies that specialize in plagiarism detection (i.e. TurnItIn, VeriCite) that would be able to detect copied text quickly and efficiently. Amazon already has easy access to thousands of books that detecting copies of chapters would be a no-brainer for them.