How can lawyers intelligently use AI?
In recent weeks, the legal press has been inundated with reports of different Artificial Intelligence (AI) tools being utilised by law firms. In case you are wondering what’s what, this post will clarify what a handful of the different systems can do, which firms are using them and the impact on legal services.
Kira Systems (Kira) and its rival RAVN are due diligence and contract review systems currently used by Clifford Chance, DLA Piper, Freshfields and Linklaters, BLP and Travers Smith. These tools analyse hundreds of documents a minute and produce reports which show different issues and trends across documents. Kira also allows lawyers to train the algorithm and to tailor the system to its own purpose.
Luminance, another speedy due diligence tool which analyses hundreds of documents a minute, is used by Slaughter and May. It also charts lawyers' progress by showing which fee earner is assigned to each task and creates a feedback loop for notes and questions.
Changing tack from due diligence: Thomson Reuters is utilising IBM Watson, a cognitive computing platform which can continuously learn through previous interactions. Reuters is aiming to use this system for tasks including deep content analysis, processing natural language, decision support and evidence-based learning. Its output is aimed at helping law firms and in-house lawyers make legal and business decisions with better data to hand, improving their efficiency and effectiveness.
Assistance in the real world
Learning about the ins and outs of these systems is all well and good, but how are they practically helping lawyers?
Currently it seems that most of these new AI systems have one primary objective: cutting the time taken for due diligence. Practically, for corporate lawyers at least, this is a broadly positive development. It reduces the number of hours spent on monotonous and repetitive document review tasks, and allows for the time saved to be put to good use in analysing the problems arising from the due diligence exercise. Trainees, who often conduct the document review side of due diligence, are likely to feel the benefits of this technology the most. More senior lawyers will be able to focus on the more engaging aspects of corporate work to which they can truly add value and expertise.
The uptake of AI is likely to increase, as demonstrated by the fact that none of the above firms used RAVN or Kira until little over a year ago. Slaughter and May now plans to use Luminance for half of its London-based M&A deal work to help speed up the normally arduous due diligence process.
However, the use of AI systems is still largely limited to the corporate law arena. It will be interesting to see what impact the systems can have on the litigation sphere; in a complex dispute, disclosure can involve tens of millions of documents and so using AI should be a logical next step. Perhaps firms’ reticence is because the courts have only just started to set guidelines for the use of AI in law.
In February, RPC won the widely publicised Pyrrho High Court ruling which approved (for the first time in the English Courts) use of predictive coding in a litigation disclosure exercise. The Pyrrho judgment was then successfully relied upon by BLP in May to obtain approval of predictive coding where another party contested its use. These favourable judgments may help to accelerate the creation and acceptance of disclosure-based AI technologies.
Earlier in the year, we considered whether technology would mean an "end to lawyers" and it would seem that the conclusion reached by David still rings true - AI is assisting lawyers to move away from one-dimensional repetitive data-intensive tasks but not yet threatening our jobs. Whilst the use of AI is growing rapidly and the number of firms willing to use it is increasing, particularly to assist in processing huge amounts of data, ultimately it does not yet have the capability to deal with the later stages of a deal or litigation.