If you want to understand how supersmart computers, aka Artificial Intelligence, might be able to fix what ails the legal system, the best place to start is a paper written over twenty five years ago.
In 1989, in their paper “The Potential of Artificial Intelligence to Solve The Crisis in Our Legal System,” (Communications of the ACM, August 1989), Donald Berman and Carol Hafner started with a relatively simple hypothetical. A guy is fighting mad at a broker who sold him a lemon of a house, and the buyer goes to talk to a lawyer about whether he has a case:
The buyer complains to his attorney that during the broker’s showing of the house the buyer asked whether “the house had water problems” and the broker replied, “Not to my knowledge.” Several weeks after acquiring the house for $150,000 a heavy rain storm resulted in significant water accumulation in the basement. The buyer discovers that the flooding occurred because the land on which his house is located has serious drainage problems. To prevent a reoccurrence will cost over $20,000. After discovering that the seller had moved away, the buyer called the broker who said that since “he knew nothing about the potential water problems he felt no obligation to contribute to the cost of the repairs.”
So the buyer asks what — to him anyway — seems like a pretty straightforward question — do I have a case against this jerk of a broker? But in response the buyer doesn’t get much from the attorney other than “well…. it’s complicated”:
After an hour’s consultation with his attorney the buyer learns that his case is “complicated,” the chances of obtaining redress “uncertain” and the cost of pursuing the claim will be at least $5,000.
Meanwhile the broker is talking to another attorney, and getting an equally wishy-washy answer:
In a neighboring law office the broker hears from his attorney that the “law is in a state of flux.” Therefore, even though the broker’s statement was truthful, the buyer stands “a chance” and that to defend a law suit may cost “about $5,000.”
In other words “well…. it’s complicated.”
So, ok, too bad for the buyer (and the broker — and anyone else expecting a straightforward answer from an attorney) but what’s this got to do with artificial intelligence?
Artificial Intelligence is a term that gets thrown around a lot to mean a lot of different things but a rough and ready definition is that it means using a computer to do something that we used to think could only be done by human mind.
In the legal system, an essential element of what an expert human mind does is try to predict what a judge will do with your case: the high hat mergers and acquisitions lawyer who has never set foot inside of a courtroom is among other things, implicitly predicting that if it ever came to a lawsuit, the fancy leaseback transaction would be held valid. The patent lawyer filing your application for a patent is, among things, implicitly predicting that 10 years from now when you are going toe to toe with a competitor claiming your patent is invalid, you are going to win.
So back to our poor buyer (or, depending on your perspective, our poor broker). Imagine a computer algorithm smart enough to give him a better answer than “well, it’s complicated.” What would it look like?