Recently, there has been some discussion out of IBM about the possibility of using Watson-like technology for legal research, litigation and discovery. While this sounds like a great idea in the abstract, in reality, it remains to be seen whether Watson is capable of such an undertaking.
For those who have no idea what I’m talking about, Watson is a room-sized computer created by IBM and named after its first president, Thomas J. Watson, that is capable of answering queries phrased in natural language. Watson became famous this past February when “he” prevailed on Jeopardy! against two of the biggest winners in the game show’s history.
According to IBM fellow Guru Rao, IBM is working towards being able to use Watson-like technology “to weed out relevant information from warehouses of data.” For lawyers, the thought of having a machine that can go through a mountain of discovery and almost instantaneously obtain the most relevant documents sounds like a dream come true. This technology could also be useful for pinpointing the perfect case or statute when engaging in legal research.
But will lawyers be willing to take Watson for his word when he is less than 100% confident in a conclusion? While Watson answered many clues correctly on Jeopardy!, he was actually less than 50% confident in many of his responses. Also, Watson’s confidence and accuracy improved with longer, wordier clues, and decreased with shorter clues with fewer words; thus, simple queries like “instances of malpractice” would be the least likely to produce reliable results.
I, for one, am looking forward to seeing where IBM is able to take this technology in order to help lawyers. Watson may be fully capable of winning a game show, but whether or not he will be a winner in the courtroom remains to be seen.