John Armour, Richard Parnham and Mari Sako have put up an interesting paper on ‘Augmented Lawyering’. The central aim is to consider what impact artificial intelligence and “associated digital technology” will have on the work of lawyers and the structure of law firms. It is an empirical study based on 52 interviews and a survey of an uncertain number of solicitors (it looks like it might be all those on the Roll) which received 353 responses. It concludes that the capital raising necessary to invest in technology is not a barrier to the application of AI in law firms (it may mean large law firms). “A central theme is that prior debate focusing on the “human versus technology” aspect of change overlooks the way in which technology is transforming the human dimensions of legal services.” The main findings are of a positive association between multidisciplinary teams and “AI–based lawtech” and that lawyers working in law firms are less likely to work in multidisciplinary teams than others they surveyed (likely, in particular, to be in-house lawyers). These findings were to be expected, of course, but the paper is interesting well beyond that.
They do a really nice job of summarising the relevance of AI (which it seems to confine to machine learning, rather than analytical (‘decision tree’ expert systems) which have a more muted presence if it is there at all in their discussion of “automated systems”. They quickly emphasise a key point, that AI is currently being deployed in, “tasks for which relevant data are readily available and in which results can be scaled, such as discovery review, contractor analytics, and legal research.” (my emphasis) They also do a neat job of summarising what natural language processing is good at and why it struggles with law (legal language being so contextual it requires a great deal of social intelligence to apply and interpret it). To those of you interested, they emphasise for areas of current application where there is scale: (E discovery/technology -assisted review; contract analytics and due diligence (with some development around risk analytics); legal research; and, billing and utilisation.
They suggest that in legal services technology first “substitutes for humans” through automated systems; then it “augments the capacity” of lawyers working with such systems; and, thirdly it generates, “new roles for humans working in MDT’s,” part of which, “involves the application of legal expertise.” It is, they conclude from their data, the absence or presence of multidisciplinary teams which facilitates or inhibits the adoption of technology. Law firms can be inhospitable places for those who are not lawyers, they note from their interviewees. Perhaps the most intriguing and important thought they have is that there is a, “strong economic rationale for traditional legal advisory firms” and that this may mean there is an occupational split between law firms that remain predominantly in this mould and other providers who are more multidisciplinary. The latter may be populated by hybrid lawyers, amongst others, who may not (I would say almost certainly do not) require the same training as admitted lawyers. In-house teams are already a sizeable chunk of the solicitors’ profession in particular, and Bill Henderson’s work in the States suggests alternative legal service providers are growing steadily, remorselessly even. To my mind the paper suffers a little from techno centrism. It is natural enough, the object of their studies is artificial intelligence and its relation with lawyers rather than legal work. They talk about their “human capital mixing the AI pipeline” although they do recognise the importance of “design schools” and “computer science”. I’d be interested to hear what the design people think of the process of product development described as one of specification, design followed by data ingestion, followed by labelling of data, followed by application. It feels quite mechanistic to me: “the pipeline nature of the exercise reveals the importance of planning and managing the process” and “it’s more like a production line kind of technology,” but that probably reflects simply their more Birds Eye view they have of “AI” in “law firms”. Henderson’s work on Legal Evolution suggests to me a more fluid, sociological process, about the construction of work and the difficulties of prising open the black box of how lawyers run cases (see here for one piece from Bill’s blog which is interesting on, of all things, metrics). Put another way, and here I’m drawing on the thinking of people like Alex Hamilton (see below), the aim feels like it’s the implementation of “digital” rather than the improvement of the product or the service that is seen as important.
Importantly though, one can begin to see traces here in their data of the ways in which multidisciplinary teams are beginning to flip that understanding. Such teams seem to succeed most strongly where digitisation is already strong within the organisation, and the paper suggests this may be associated primarily with, “increasing revenues and keeping costs down.” So it feels like a technology-efficiency dynamic at play. And such teams seem to major on, “data science, project management, and design thinking”. Perhaps it is the latter that provides the best hope of rounding the edges off, replacing even, techno-centrism.
In sum, law firms have lawyers as the solution to your problems. Law Tech types have “law Tech” as a solution to your (large-scale) problems. The paper’s key point is they see law firms as keeping control of legal problems because of their very contextual, bespoke nature. Lawtech, or AI, cannot (yet or for the foreseeable future) break those down. The interesting question for me is what does a third group have? That might be who our eyes should be on because they do not privilege one type of knowledge or another, if enough clients start to see them as competent, perhaps they can challenge. In-house legal teams, for instance, are not only or mainly interesting because they use law Tech; they are interesting because they have a different way of lawyering and a different idea service. Or at least some of them do. Some external legal providers too. The most worked through model of it I have seen comes from Alex Hamilton of Radiant Law and he discusses it in this post that I featured here as part of my interest in Cybernetic Lawyering.