3 ms·

I was pondering this further and it would actually be conceivable with a bit of natural language processing and machine learning to build a 'verdict machine', m

I was pondering this further and it would actually be conceivable with a bit of natural language processing and machine learning to build a 'verdict machine', maybe a Watson type of thing. You would input legal documents, contracts and laws to 'program' the machine, It would read these legal documents and compile them into legal tests that depend on case facts. Then when you needed to judge a case, you would enter the facts or fact probabilities about the case in the machine. It would then generate a verdict and tell you which legal tests passed or failed to arrive at the verdict. Even it it was just used as support to help courts untangle cases or for law student as an educational tool, it could be a useful machine. Here is the kicker: All contracts and laws are actually software code for this machine.

jeremyjh 14y ago
You've certainly got lots of data to train it with. And you can test it by simply testing its predictive power. But instead of selling it to the court system I'd sell it as a service to law firms and/or use it in prediction markets. This would ensure that it would be funded and refined until everyone simply accepted its outcomes as a foregone conclusion.