Automation in the Legal Market

Tom Davenport is, among other things, a distinguished professor at Babson College, a research fellow at the MIT Center for Digital Business, and director of research at the International Institute for Analytics. He is recognized as one of the pioneers in the field of knowledge management (KM). In 2000 he co-authored, along with Larry Prusack, one of the early books in the field of KM, Working Knowledge: How Organizations Manage What They Know. He is acknowledged as one of the most trusted consultants and the third leading business-strategy analyst (just behind Peter Drucker and Tom Friedman) by Optimize Magazine. (Read more about him in the bio on his website.) So anything he writes or talks about is immediately of interest to those of us in the KM field.

Two recent articles, one that he wrote and one where he’s interviewed, provide some insight into the technological changes that are making themselves felt, among other places, in the legal market. (I deliberately use the term “legal market” rather than “legal profession,” to try and highlight the fact that the forces at play in other markets are already being felt amongst lawyers and law firms. Simply calling this a “profession” does not somehow keep those forces at bay!)

His article, from February 2015, is on the CIO Journal blog of the Wall Street Journal and is entitled Augmentation or Automation? The interview with him, also from February 2015, can be found on the site of APQC, a member-based non-profit organization and one of the world’s leading proponents of business benchmarking, best practices, and knowledge management research, as part of their Big Thinkers, Big Ideas series of interviews, which includes one with Larry Prusack. (If you’re interested, the APQC site offers research and expertise in five core areas: business excellence, financial management, human capital management, knowledge management, and supply chain management.)

In the interview, Mr. Davenport notes that he and Julia Kirby are working on a new book about how knowledge workers can be augmented by smart machines — or be automated out of a job, which is the topic he treats in his WSJ blog posting.

He defines automation as “the takeover of jobs and tasks by machines” and says that we’re into the third era of automation, which now involves knowledge workers. “We’ve always thought that knowledge work was too unstructured and judgment-oriented to be automated. But if you look around, some of the highest-valued knowledge work — jobs like doctors, lawyers, and accountants — can be increasingly automated these days.”

He notes that, “Most of what knowledge workers uniquely do is using their judgment to make better decisions. [But] disintermediation is happening to knowledge workers as well… [So] if you aren’t making effective use of the knowledge you’ve accumulated, then you have a pretty good chance of losing your job.”

He goes on to discuss five steps that humans can take to augment smart machines. He defines augmentation as “a situation in which humans and computers combine to create effective and efficient outcomes.” Each involves a type of specialization and, although he does not specifically say so, involves shedding many of the people currently doing the tasks that will be displaced by machines.

In his WSJ blog post, he talks about a “smart system” implemented at a number of Boston hospitals run by Partners Healthcare as an aid to processing medical orders (drugs, tests, referrals, etc.). The system “checks to see if the order is consistent with what it thinks is best medical practice. If not, it asks the physician if it wants to change the order, although the final decision is up to the doctor. When the system was implemented in two major hospitals at Partners, it led to a 55% reduction in serious medication errors.”

The most obvious example of automation in the professional services field is the way that the Memorial Sloan Kettering Cancer Center in Boston is partnering with IBM Watson to address cancer diagnoses. Tim Knight’s Slaw posting entitled Deep Learning Algorithms and the “Machine Learning Revolution” talks about this partnership and links to a fascinating TEDx chat by Jeremy Howard entitled “The wonderful and terrifying implications of computers that can learn.”

I very much had that TEDx talk in mind during a recent demo of the ROSS legal research tool, which, it notes prominently, is “Powered by IBM Watson.” (You can find out more about ROSS here and here.) While it is a team out of the University of Toronto that won the right to develop a legal application for Watson, the intention is that this will be a tool for multiple legal markets, including all the common law jurisdictions.

The demo we saw was disappointing: the sample search that was used on a fairly straightforward employment law question, namely whether a former employee can leave and start up a competitive company, dredged up a 1967 SCC decision on the issue. I understand that there has been much more recent case law on this topic. In the discussion afterwards amongst other KM practitioners, the sense was that the ROSS team has so far been unable to secure the rights to legislation, case law, and legal publications that will be necessary to make it run properly.

It was, for example, clearly in Sloan Kettering’s interest to give Watson as much information as possible, including millions of pages of text from academic journals and clinical trials, as well as hundreds of thousands of patient records, including test results, x-rays, etc. Watson can, therefore, suggest possible treatment and offer percentage likelihoods of success and make diagnoses that are more accurate than those of the most experienced oncologists.

Right now, ROSS has apparently been unable to enter into an arrangement with any of the large legal publishers, since the publishers have no interest in turning over their intellectual capital to a competitor. Similarly, I understand that repositories such as CanLII have been architected to prevent commercial services such as ROSS from indexing their content.

But one assumes that, sooner or later, the ROSS investors will find a way to get their hands on the appropriate content. And when they do, there is no reason to believe that the results will be any less impressive than what Watson can already achieve in the area of cancer treatment.

So, where does that leave most practising lawyers? Will they be like the 144,000 Kodak employees at the peak of its operations, only a few years before it went out of business, victims of the changes of the new digital world?

Mr. Davenport would argue that the path for knowledge workers is through augmentation, in other words by partnering with the machine to provide something “more” than the “mere machine” can provide.

In the APQC interview, Mr. Davenport talks about “five steps humans can take to augment smart machines.” My guess is that smart lawyers will choose one of those steps and create niches that will allow them to flourish. However, recent studies of the effect of digitization on the work force, including Erik Brynjolfsson and Andrew McAfee’s book The Second Machine Age, do not leave me hopeful that lawyers will fare much better than the ex-Kodak employees. Time will tell.