The next legal era: Artificial intelligence and cognitive computing
We are on the cusp of a new era in computer technology. What could possibly connect rapid automated legal decisions, systematized stack ranked medical diagnosis, and never created scrumptious food recipes? At the IBM Cloud Innovation Forum last week in Dana Point, CA, Mike Rhodin, senior vice president of IBM Watson, connected the three with a provocative overview of what our future in legal entails.
Here is some brief history. The computer is entering its third era. In the beginning, we focused around the computation of numbers called the Tabulation Era. Simple calculations ruled this phase. After decades passed we moved into the use of magnetic tapes and programmed storage of information. The central piece surrounding this period was programming and so is known as the Programmatic Era. Currently we find ourselves in this phase but edging into the third known as the Learning Systems Era.
Our current era is in a troubled state. In 2020, all data available will surpass 40 zettabytes (40 trillion GBs). The slope on a graph which represents this change from today is nearly straight up. For comparison, in 2009 the entire World Wide Web was estimated to contain close to 500 exabytes which is about one half a zettabyte. Every electronic act, button pushed, video or picture taken, word written, is a magnification of significant proportion. However, where data storage multiplies is as we create and store meta data and learned connections from the base data.
Why is data exploding? When Watson (IBM’s Artificial Intelligence) competed against Ken Jennings and Brad Rutter on Jeopardy, IBM gave it a brain (offline hard drive) of all the information they thought was pertinent to win. Initially they found it was about 10 gigs of information. However when the engineers extracted the meta data from the 10 gigs using every algorithm imaginable, the storage ballooned 10 times to 100 gigs. Lastly they applied other computer assisted cognitive learning against the data and the storage grew by 100 times. This is one reason the amount of data is exploding as we move toward 2020. The data we create about data is actually more voluminous than the original data set, and the learning’s from that are even greater. The volume of data is at a tipping point. We are nearly at capacity to understand it as humans.
Rhodin says we make decisions through four phases: we observe, interpret, evaluate and then decide. Watson now has the ability to tap into this same process. This was made evident when doctors saw Watson on Jeopardy ranking top answers based on evidence and supporting hypotheses. These data points can help form decisions.
Expertise matters more than ever before and with the amount of data coming online it is increasingly challenging for any one human in a field such as law to have mastery. Cognitive computing enhances, accelerates and scales human expertise so that the attorney can wrap their head around the patterns and make intelligent decisions. The attorney has to be especially aware of this as they are presented with much larger volumes of discovery data than ever conceived.
In a very unique example of Watson’s prowess, it digested all of Bon Appetite’s recipes. The system was trained in underlining chemical combinations thus bringing never tried together favors with very different blends. It connected the dots on things never created and spit out several recipes. The food was launched in a food truck at the festival South By Southwest (SXSW) in Austin, TX, where it produced huge buzz with lines extending three blocks. Chef Watson’s cookbook of 100 recipes is coming out in the coming months.
Ultimately what Watson is doing is exploring the whitespace, attempting to connect the dots into what has not been found. As we are inundated by data, cognitive computing will help experts sift through the data and retain their expertise. It is clear that this is the future of many industries including the legal sphere. Watson is in beta trials for medical and life sciences, law enforcement, financial services and legal.
Dear Joe,
can you explain a little more about the automated legal decisions ? Do you know some research/project running on this now ? I am an attorney and librarian, in Rio de Janeiro, Brazil and I´m interested on this.
You can find me on Linkedin:
Renata Marques
Attorney – Legal Knowledge Management and Business Intelligence specialist. at Law Office
Best Regards,
Renata Marques.
Hi Renata,
Thanks for your note! That is a great question. They did not delve too far into specifics surrounding their beta program; however, I have some possible ideas.
1. Dispute Resolution: The two engaged in a dispute could submit all of their information into a form with narrative and Watson could compare that to previous like disputes – helping either automatically reach a decision or offering up a summary for the arbitrator.
2. Patents: Watson could digest the entire array of current patents. All new patents could be vetted by Watson and identify where a patent application is similar to others.
3. eDiscovery: I see Watson sifting through massive quantities of eDiscovery and not only being able to find relevant connections expected, but finding new connections not yet established.
4. Judging: This is a leap, and somewhat futuristic, but I do see the possibility of a Judge Watson, which could have such a vast array of knowledge and resources coupled with its logic that it could begin making decisions on cases involving legal code. That is decide on more cut and dry cases where less reason was necessary, until we began to trust Judge Watson and then more complicated cases could be brought.
Truly amazing possibilities here.
Thanks!
-Joe