If he published anonymously surely with machine learning these books could be identified for prose similarity?

I imagine if Google has OCRd nearly all the books this would be an interesting problem for someone there.

Unless they’ve already done it.