Mining novels

This is a great story by John Markoff in the NY Times. I'm not a lawyer though, or I'm sure I'd see it differently. Legal discovery has been automated, and the sophistication of the software is mind-boggling. The sweetest part is the Enron Corpus, a public database of Enron email and documents seized by the government during the criminal proceedings, which has proven invaluable to researchers trying to understand corporate language and social networks. Future Enrons will be plagued and prosecuted, not by lawyers, because there won't be any, but by software developed with the Enron Corpus.

A novel could be written with this data-mining software. There must be a million stories in the Enron Corpus. Obviously plenty of slimy, criminal tales, but also romance, misunderstandings, jokes, lies. Happy families, miserable ones, duck hunting, moving children to college, crazy drives across country to look up people who died years ago. No more writing novels - we'd just start our software and make coffee. It's 100% accurate, while humans are barely more accurate than flipping a coin.