CLE Conference.  Final thoughts: doing and undoing

by Paul Maharg on 20/06/2017

In the FT back in March, John Gapper wrote an interesting article on why Standard Life offered a whopping £3.8bn for Aberdeen Asset Management.[1]  It wasn’t a merger based on strength on either side (Standard Life acquiring Aberdeen’s niche skills; Aberden accessing Standard Life’s strength in developed markets) but on joint weakness.  That weakness, Gapper argued, is one that the entire investment management industry is currently facing, namely that algorithms are out-performing many active fund managers.  There is a place for niche fund management argued Gapper, but ‘formulas are perfectly good at doing the predictable stuff, and often better’.  And sure enough, soon after Standard Life and Aberdeen completed the negotiations 800 redundancies were announced – a total annual saving for the two Scottish companies of over £200M.  Shareholders voted yesterday for the merger to go ahead.  Given the cost to investors of active-managed funds (largely the salary costs of managers and related staff) as against indexed and exchange-traded funds, and the steady drift of investment to the latter, there was little doubt which way the vote would go.

It’s not only investment management that’s facing such change: every aspect of financial work is under the same relentless pressure.  Indeed pretty much every industry is affected by the digital revolution we’re living through.  A washing machine is now a computer that washes clothes; a car with or without a driver is a computer that takes you places.  Increasingly, houses are becoming computers in which we live.  In late capitalist societies digital shapes employment that’s on offer in the marketplace in ways often invisible to the players involved: the number of jobs, their activities, their burdens and rewards.  Digital shapes work, creates opportunities and huge wealth but devastates lives and careers as well.

Most lawyers now realise this – it’s been slow, but the evidence is there.  Computational law – that corner of the extensive field of legal informatics that deals with the automation of legal analysis and machine learning – is growing apace.  See for example the projects of CodeX at Stanford, the work of scholars such as Daniel Martin Katz, decision-support systems that use computational learning, and engines such as (in no order other than alphabetic) Casetext, Kira, LawGeex, Legal Analytics, RAVN.  These and many other examples point to common analytical tools that apply to both numeric and linguistic domains – the unbundling of processes and tasks, automation and decision-making.

And law schools and legal education?  Ah, we’re different, you see, we deal in legal analysis and justice, not the arithmetic of the market; in legal reasoning not financial results or the data-driven models of legal practice.  So we like to think.  Meanwhile our students pass through cultures of education and curriculum practices & content that in many quarters have changed only superficially since the 1950s, and graduate through assessment regimes many of which wouldn’t look out of place in the 1870s.  Then they face the chaos and kettledrums, the professional process of doing and the undoing of much of what they’ve learned from us in law school, no matter what profession they enter; and they learn the bitter lessons of marketplace ethics and economics, about which we taught them so little.  The common analytical tools I mention above, for instance  – where do they appear in a curriculum?  And yet over 40 years ago educationalists such as Lee Shulman were already applying such research to educational theory and practice.[2]  What passes for digital literacy and its processes in our curricula?  Disintermediation, as I argued recently, is a process that applies not just to industries but to legal knowledge itself in the last millennium of legal scholarship, yet it is a process of doing and undoing that we scarcely discuss with our students.  As I said in Transforming Legal Education, we’ve yet to attain in our theory and practice in legal education what Ernst Cassirer called ‘mature constructivism’, namely the self-reflexive view of the development of technology within the history and culture of the domain.[3]

Just why that’s so is too complex to go into here; but the end result is the same: we set up wasteful processes of doing and undoing, the ceaseless antinomies whereby students are first socialized into a view of law and technology that requires to be unpicked in lives and social practices beyond law school.  Surely we can imagine better models than this?  Katz points to a useful thought experiment – the MIT School of Law[4]  A decade ago in Transforming Legal Education I described a future University of Scotland Law School where students lived and learned in AR, where they learned how to negotiate the process of professional relationships, and learned justice for a global world through transactions, assisted throughout by avatars.  In such environments doing and undoing is not a process of unpicking knowledge but a process of becoming through experiential forming and reforming in saturated technological environments.

We need such new models more than ever.  In his FT article Gapper references Michael Lewis’s book, The Undoing Project, which describes the work of the psychologists Daniel Kahneman and Amos Tversky, and in particular their studies of what they called ‘low validity environments’ – domains of human activity prone to uncertainty and unpredictability’, just the fields one might think where machines might have difficulty and where human judgment and intuition would predominate.  Not so.  In one study the prediction of cancerous tissue by radiologists from the evidence in x-rays, for instance, was outdone by simple algorithms.  Does this mean that radiologists can be replaced by robots?  That’s the wrong conclusion: some professional tasks can be performed by algorithm more successfully than by humans in all professions, but not all.  Too often we stop at that point to observe the rise of the machine – ‘[Kahneman and Tversky] showed clearly half a century ago that would happen, and now it has’, as Gapper says of the Standard Life merger.

The dominance of the robot algorithm is not the only conclusion we can reach of these types of study, though.  Expertise is a highly complex collection of skills and knowledge, and even in their own fields experts such as radiologists are more expert in some practices and domains than in others, sometimes just by the nature of what they do and how often they do it.  The reasons why radiologists’ judgments were wrong were complex, and not least in that complexity was the nature of their profession and the extent to which they received feedback upon their decisions.  Kahneman puts it well:

Among medical specialties, anesthesiologists benefit from good feedback, because the effects of their actions are likely to be quickly evident.  In contrast, radiologists obtain little information about the accuracy of the diagnoses they make and about the pathologies they fail to detect.  Anesthesiologists are therefore in a better position to develop useful intuitive skills.  If an anesthesiologist says, “I have a feeling something is wrong,” everyone in the operating room should be prepared for an emergency.[5]

So in legal education, let’s bury the tired notion that we’re teaching students how to think like lawyers and ask instead how are we helping them to learn how to decide and judge, how are we helping them to shape and live their lives?  And does our education put them in the position of being radiologists or anaesthetists?  There lies our expertise.

Amongst many themes, that was one that was raised consistently in the papers presented at the Centre for Legal Education conference, and in the many interesting discussions that arose from them.  What was memorable about the event, apart from the lovely sense of collegiality, was the collective feeling that we were engage in exploring educational innovations that challenged us to think hard about our educational practices, to rethink them, to do and undo in the most creative sense what we did with our students.  The liveblog postings that Pamela and I put up here and on the CLE blog catch just some of the ideas and practices we experienced.  My thanks to the centre director Jane Ching and all her team for a great conference.  If you happened to miss it, see you there next time…

  1. [1]John Gapper, Technology outsmarts the human investor’, Financial Times, 9 March 2017, p.11
  2. [2]See eg Shulman, L.S., Elstein, A.S. (1975).  Studies of problem solving, judgment, and decision making: implications for educational research.  Review of Research in Education, 3, 1, 3-42.
  3. [3]Cassirer, E. (1946).  Language and Myth.  Translated by Susanne K. Langer, New York, Dover Publications, Inc.
  4. [4]Katz, D.M. (2014).  The MIT School of Law?  A perspective on legal education in the 21st century.  University of Illinois Law Review, 5, 101-42.
  5. [5]Kahneman, D. (2011).  Thinking, Fast and Slow.  London, Allen Lane, 242.

{ 0 comments… add one now }

Leave a Comment

{ 1 trackback }

Previous post:

Follow me on Academia.edu