Robert Morley, chief operating officer at Excello Law, examines the use of AI in the legal industry.
Fear of intelligent machines makes news.
Online, a clear winner has emerged from this unlikely trio of scary topics: AI returns nearly two billion results in Google search, whereas ‘The Donald’ scores a more modest 368 million and Brexit a mere 108 million.
Among articles and blogs about the future of law firms, AI has similarly dominated attention as the number one theme. Hyperbole has flourished. Underpinned by dystopian visions of lawyers being replaced by robots and the growing ranks of tech companies dedicated to replacing the human element from much of day-to-day legal practice, the future for the lawyer has seemed bleak.
Of course there is nothing new in the idea. Professor Richard Susskind has successfully cast himself in the role of Cassandra for the legal profession over more than 30 years, arguing that their numbers will shrink dramatically as the ever increasing power of computers will inevitably lead to AI undertaking most of the routine work.
It has been an entertaining and lucrative narrative, carefully re-crafted with each stage of technological development in a series of very readable accounts. The only problem is that the fundamental premise of his argument – that AI based technology will make most lawyers redundant – is false. In the 1980s, Susskind’s original algorithmic rule-based programming approach was abandoned by the AI community as a cul-de-sac. It hit a dead end.
Meanwhile the approach has moved on apace. By combining a bundle of different technologies, such as machine learning underpinned by neural networks, powered by vast arrays of parallel computers, we now have the ability to rapidly analyse very big datasets for patterns.
Modern AI is increasingly able to perform complex, repetitive analytical tasks much faster than humans. Law firms are now using multiple. different AI applications, powered by systems such as IBM’s Watson – celebrated for winning the US game show Jeopardy.
IBM’s Deep Blue has also defeated world Chess champions while more recently Google AI beat Lee Sedol one of the world’s best GO players.
AI applications like Watson improve and enhance legal research functions, enabling lawyers to provide dramatically faster and more comprehensive analysis of contracts and precedents.
A landmark case last year, Pyrrho Investments v MWB Property saw the first approval by an English court concerning the use of predictive coding in document review: an AI system was trained to identify and classify relevant documents within large volumes of data. But like every other AI application, predictive coding is a tool. It does not replace lawyers, but instead frees up their time for other tasks. It adds to efficiency in the same way that the personal computer, the mobile phone and the internet have all done within the lifetime of many senior lawyers still working today. Indeed, AI is becoming an integral addition to each of these existing technologies.
As AI became mainstream in 2016, even the magic circle firms wedded themselves to forging deals with AI tech providers: Clifford Chance to Canadian software provider Kira Systems, Linklaters to London-based AI service provider RAVN. And so on.
But the number of lawyers employed by large commercial firms, and in particular, the number of trainees taken on – the best indicator of future size – has not decreased. Instead, since the post-crisis low of 2009/10, the number of lawyers and new trainees has increased.
As big commercial law firms begin to embrace the opportunities that AI presents, they are not planning to reduce their headcount. If anything, the reverse appears to be the case. True, there are variations: Clifford Chance reduced training contract (TC) places from 100 to 80 last year, but this was compensated for by White & Case, which simultaneously increased its TC numbers from 30 to 50. Overall, training contract numbers rose by 9% in 2016 from 5,000 to 5,450, according to the most recently available figures.
Putting aside fear and hyperbole, the latest machine learning techniques driven by high capability parallel processors, has many exciting possibilities. As with the medical profession, where AI tools are already being used for early cancer diagnosis, lawyers will increasingly use AI as a diagnostic tool, to support their work in myriad ways. But however fast and versatile, AI will remain a tool, incapable of creative, independent judgement – the essential elements in the job of a lawyer.
However clever our machines become, they can never be human. No algorithm exists to replace us. Understanding when a situation is subtly but importantly different from before, and knowing how to handle this new situation, is a distinctly human characteristic. While current AI approaches have successfully taken us from data processing, to the extraction of information and now knowledge, the uniquely human characteristics of wisdom will not emerge from them.
As inspiring and exciting as the current generation of AI research may be for what it can do in saving time and minimising laborious legal tasks, it will remain a useful tool. Self-aware artificially intelligent robots have filled the pages of science fiction novels for nearly a century since the word was first coined in 1920. Over the coming century, they will increasingly help lawyers to perform their job in many different ways.