Monday, March 23, 2015

A Tale of Two Conversations: Is what lawyers do really special enough to be an exception to automation?

Over the past few weeks, I  attended a couple of contrasting events that reflect the “battle lines”, if I may call them that, on the potential for technology to disrupt the way we currently practice and teach law. 

The first was a session of “legal rounds” held by Legal OnRamp at which I was an invited guest (via call-in through a webex connection).  This company is one of the new breed of legal service providers, with a mission to offer “Legal Intelligence Solutions that leverage expert people, best practice process, and state-of-the-art systems to support excellence in internal legal functions and tackle some of the most complex projects faced by legal teams.” See  The conversation was lively, with lots of youthful enthusiasm and optimism. A good portion of the time was devoted to a serious discussion of how to allocate what we think of as legal work between people and machines, using as a starting point the concept of “heteromation” discussed in a September 2, 2014 posting on this site. 

A week later, Stanford’s CodeX Center for Legal Informatics and the Palo Alto Area Bar Association sponsored a panel on The Law Firm of the Future, and I was fortunate to be a member of the audience.  See the poster at and the video of the proceedings at  Most of the participants on this panel – largely established practitioners – took the position that there was something inherently special about what lawyers do that will prevent the successful automation of those processes.  The one exception went, if anything, too far the other direction – although his starting point seemed valid to me. Such predominant conservatism is unusual for a CodeX event.

Given the nature of this blog, it is to be expected that I agreed with the speakers in the first session more than the second.  I came away from the Stanford panel a bit depressed.  The arrogance of our profession can be so tiresome.  Just because smart people can become very good at what they do does not mean a) that technology can’t improve/substitute for the process or b) that what they do is in fact the best way to accomplish the goal.  The pattern recognition skills that several of the speakers referred to as uniquely human are in fact likely to be far easier to automate than truly hard feature extraction problems like voice recognition software.   At this point, making progress on many legal automation targets is just a matter of envisioning the next steps and applying resources.  If legal tech startup funding continues to ramp up, the resource part may not be that much of a problem going forward.

Law is in the early stages of a relatively common, if often rocky, story: the passing of the importance of what has been a highly valued form of competence.  To illustrate this, think about the role of horsemanship in transportation.  Horsemanship was a terrific skill when horses were the best mode of land-based travel humans could access.  Really competent riders were at a premium and justifiably honored for their accomplishments.  When automobiles came along, however, suddenly horsemanship became largely irrelevant.  The analogy doesn’t map point for point on the changes we’ll see in the law, but the overall conclusions are useful – being really good at what has been in demand doesn’t make you irreplaceable.  In some ways, it just sets you up for a bigger fall.

| Permalink


Post a comment