Look left: the machines are coming
Lawyers like nothing more than a good argument. That’s one of the attractions of the job for us.
And the arguments are raging over the role of technology in law and driving some of us into two camps:
- what I might call the Canutes, who are looking at the rising tide and saying that it will stop before it gets to them; and
- the Chicken Lickens, who are racing around telling all who hear them that the sky is falling in.
It’s a debate Mark Gould and I played out across two editions of Legal IT Today inJune and September last year. I was in the Chicken camp; Mark in the Canutes. The latest exposition of the Canute theory of the world was by Richard Tromans last week.
The Canutes reason like this:
law is a cerebral process, but it is not just process. The best lawyers use intuition, innovation and imagination to solve client problems.
What’s more, clients don’t just buy lawyers on the basis of cold, hard reason. There’s an emotional, irrational element to buying decisions, whether it be simply getting on better with the person pitching to you, or simply ‘trusting’ the person whom you know better than the one you don’t.
One version of the Canute argument also reasons that there will actually be a positive impact on gender diversity in lawyers’ senior ranks: the machines will be able to replicate only the logical left-brain activities (statistically more common in men) than the creative right-hand brains (more common in women), with the result that women and their intuitive skills will become more highly valued and will rise through the ranks more easily than men. (The left brain/right brain theory has been generally debunked as a myth, but there does seem to be some evidence of different wiring for men’s and women’s brains, with the latter more creative and intuitive as a result.)
On the other hand, the sky-falling-in Chickens (like me) say that there are a number of factors at play here. First, the machines are getting undeniably brighter, with human level machine intelligence (HMLI) inevitable. I heard Nick Bostrom explain last week that the overall predictions currently show a 10% chance that this will be achieved by 2022, a 50% chance by 2040 and 90% by 2075. (That was in the context of machines presenting an existential threat to all human life, and not just lawyers. One for another time maybe. Let’s just try to save the lawyers first.)
The Chickens say that:
the proportion of creative thinking and intuition needed in law is relatively small. Not many minutes of creative thinking are really needed in matters taking many days or weeks of time to complete.
And isn’t it the case, really, that these activities are seen as ‘creative’ simply because we are having to reason on likely outcomes in the absence of quantifiable data? Once law firm data analytics progress to the level where lawyers are able to mine past transactions and case outcomes to assess risk more effectively, then the space left for hunches and creativity is reduced to virtually nothing.
If courts start to take decisions driven on data (including data as to the likely truthfulness of witnesses, based on observed physiological reactions when giving evidence), the days of judges giving bad decisions because they are hungry will be a distant, painful memory.
Last but by means least is the fact that more buying decisions are already being based on value – maybe not absolute price, but overall cost (although many are already purely on price). The rise of the procurement function in corporates is undeniable, and law is increasingly seen as a commodity to be scored, rated and ranked like any other. When combined with data driven, quantified risk assessment, buying on sentiment will seem curiously quaint and quasi-astrological pretty soon.
The crux of the issue is the obverse of David Maister’s classic Rocket Science to Commodity progression for professional services. His argument is that an innovative service that is initially seen as complex and unique will become ubiquitous and commoditised over time as it is replicated and ‘drifts to the left’ down the complexity scale.
The reality today is that machines are driving fast from the left to the right, moving up the value scale from commodity onwards and taking over large volumes of work previously carried out by qualified lawyers. First, they are allowing those qualified lawyers to be replaced by less or unqualified staff. Second, they will start to replace humans entirely for those tasks, processes and decisions.
Already at Pinsent Masons we are going beyond the theory and are mixing AI with our core technologies to assist on real matters.
- We are using our own AI legal rules engine to map and control matter processes, and are proving how this technology can be extended to read and categorise contractual clauses (indemnity, assignment of IP etc) to be analysed and reported against by RAG matrices.
- We are looking at the data we hold, through bespoke matter management systems, and highlighting for clients the trends we can see in the reasons why they are coming to us, so that we can work with them to prevent those issues arising in the first place.
- And we are looking at how we can analyse data across our disputes practices to identify which factors applied to particular cases and what patterns there are in the outcomes achieved, so that we can provide early assessment of likely results when similar factors are present.
The more we look, the more opportunity we can see to use machines to help solve – or avoid – problems for clients. And the space left for real-life human lawyers risks being increasingly squeezed and competitive as a result.
13 Mar 2015