PREDICTIVE ANALYTICS HELP LAWYERS
MAKE PRECISE DECISIONS
by Don Philbin of Picture It Settled
Cheap storage and fast computers have accelerated
information production. We create more
information every two days than we did from the
dawn of civilization through 2003, according to
Google's Eric Schmidt. Legal is no exception. It took
U.S. appellate judges 69 years to fill the Second
Series of the Federal Reporter. It will take less than
half that time to fill the third series, even as official
publication diminishes.
Thanks to much faster machines and distributed
computing, learning systems can mine data in new
and exciting ways. In his new book, "Tomorrow's
Lawyers," Richard Susskind notes by 2020 the
average desktop will have the processing power of
the human brain; by 2050, it will have more than
humanity combined. And it's not just to search
and retrieve static information anymore — IBM's
Watson beat the two best "Jeopardy!" contestants
two years ago using machine learning, information
retrieval and advanced natural language
understanding.
But don't wait until 2020 or 2050! Computers
already crawl the entire federal court docketing
system (PACER) daily looking for patent documents,
so practitioners can determine whether to try or
settle their cases. Predictive analytics also scour
company share prices to help value securities class
actions. Other systems analyze billions of dollars in
legal spending to help project how much a matter
will cost. Learning systems use patent-pending
algorithms and deep data on settlement moves
in thousands of cases to predict accurately where
parties will settle based on their early moves.
These are exciting times for legal technology. It
doesn't replace lawyers; it gives them a sharper
focus. Human judgment with advanced analytics
accessing lots of data is a powerful combination.
