As a lawyer, you probably hear a lot about new technologies, but don’t have time to sit and read through pages of technical information to understand if or when they’ll be relevant to your practice.
In this series, we will cover some of the major technical innovations within the legal industry from a lawyer’s perspective. We’ll tell you not just how they work, but the implications for practicing lawyers now and in the future.
Our next topic: Artificial Intelligence.
What is Artificial Intelligence?
Standard software is only able to follow instructions – instructions found within its code.
Artificially intelligent software is designed to mimic ‘human intelligence’, most specifically in terms of learning, planning and problem solving.
‘Machine learning’ is integral to this process. Just as with human intelligence, artificial intelligence develops through learning. However, in this case datasets are provided to the machine. The machine goes through an initial ‘training’ period – where it tries to complete its task and is given feedback on how well it has done. Once commercially viable, the artificial intelligence algorithm continues to incorporate feedback from data gathered, making it increasingly more accurate.
For example, the algorithm within a self-driving car needs to identify whether it should pass through a traffic light or not. If the traffic light is red, amber or green this is relatively simple. If the traffic light is broken, or is flashing, or the coloured glass has been smashed and so it is displaying a yellow light in the red-light position, the algorithm also needs to know what to do. Multiply this by all the other repetitive decisions made by a driver every minute on the road. Something that seems intuitive to a human can require a lot of machine learning for an algorithm.
Where Is Artificial Intelligence Being Used in the Legal Industry Today?
One of the key areas in which artificial intelligence is currently used is in eDiscovery. AI makes search more effective by learning about what human reviewers are looking for and providing what it believes is the most relevant information, rather than traditional Boolean searching. It makes eDiscovery quicker, cheaper and much more effective.
eDiscovery tools vary in their level of sophistication but are now relatively widespread in their use by firms engaging in litigation. The worldwide market is expected to grow to $10.7 Billion in 2018. eDiscovery software is now available for SMEs as well as enterprise firms, often on a ‘pay per gigabyte’ basis.
AI has been developed to answer legal questions, highlighting relevant information from published and unpublished case law.
AI can be used in several areas within this. The first is to understand lawyers’ questions. Boolean searching can easily miss data if you are not matching exact terms. Natural language processing, and area of AI, works to understand what information humans are looking for when they ask a question, and providing relevant material.
AI solutions can also be used to analyse a brief of other legal documents. It can provide information about cases cited within the document that highlight the most relevant factors such as cases which have received negative treatments. It can highlight the most relevant paragraphs, to aid in skim-reading. It can also look for similar cases in the same area of law.
Again, AI is used to analyse the information so that lawyers receive only the most relevant information. This saves lawyers significant amounts of time and resources.
Some free tools are available, but comprehensive offerings are primarily available with an enterprise price tag at the moment.
Recently, software provider LawGeex challenged 20 experienced lawyers to a competition – reading an NDA. LawGeex software took 24 seconds and had 94% accuracy. The 20 lawyers took an average of 92 minutes, with an average 85% accuracy.
In addition to fast and accurate identification of potential issues, AI-supported contract reading can also help companies to compare contracts across multiple business divisions, and to identify potential issues within contracts within seconds. For example, a company could specify clauses they always approve, those they always reject and grey areas. Once the software has received this information, it can present companies with an analysis of areas that need a human lawyer to look it.
This software is currently primarily aimed at the in-house market – where it has a clear use case, and large volumes of similar use cases can cross lawyers’ desks.
However, as this technology grows we expect it to cross into enterprise law firms, and eventually SMEs.
What is the Future of Artificial Intelligence in the Legal Industry?
Clear use cases for artificial intelligence have already been identified, and there has been a recent growth of legal technology start-ups building on artificial intelligence technology. Expect these applications to become better and for the number of products to multiply.
Whilst this article has primarily focussed on AI applications, as general artificial intelligence tools improve, lawyers should expect AI to become a component of everyday applications as well. For example, case management software should begin to incorporate elements of AI, and some (including Thread Legal) already do. In these instances, AI will be used to enhance automation and document management features – so will blend into the interface and become part of the user experience.
What This Means for Lawyers
Although, of course, there are as many opinions as there are people, the common consensus is that artificial intelligence will not replace the need for lawyers any time soon. What AI will primarily do is reduce the time that legal teams spend reviewing data – whether that’s contracts, discovery data, case law or other documents. Lawyers will then spend more time on higher value work that needs human involvement, primarily the analysis and judgement calls. An AI software may identify potential risk factors in a contract much more quickly than a human could – but a decision on whether to challenge the terms, or to accept them in order to avoid the deal breaking down, is a judgement call that lawyers and their clients need to make.
In terms of adoption times, there are several AI software solutions on the market that are fully developed products suitable for adoption by firms of all sizes, and in-house legal teams. However, there has been much market discussion over why these technologies are not being adopted as quickly as expected.
The general expectation is that in-house legal teams will be the first users of AI legal technology. For in-house lawyers, efficient use of resources is a key metric – and it is therefore easy to see clear return on investment. However, the initial investment in, training, and ongoing IT support of an AI product can make adoption difficult for busy teams. Whilst use is increasing, it may be a couple of years before it becomes widespread enough to reach a critical mass.
With legal firms, the adoption difficulty is that billing structures can make adopting AI seem counter-intuitive in many cases. A reduction in hours available to be billed has obvious implications – equally, a flat fee that is designed to cover weeks of work will look excessive when weeks of manual review can now be done in seconds.
Several suggestions have been made as to how AI adoption will take place. One is that it will take place in response to pressure by in-house legal teams on their outside counsel. As in-house teams become more efficient through the use of AI, they will expect outside counsel to do the same and reduce their bills accordingly, and adoption will spread from there. Another suggestion is that a new breed of young firms will arise, which automate as much as possible through technology, and grow through volume of work. Enterprise firms, and then eventually SMEs, will have to adopt AI technologies in response to the changing market conditions that these firms will create. Adoption rates will probably vary greatly by region and area of law.
However it happens, the truth is the artificial intelligence is here to stay in the legal industry. Viable technology is already available – it is up to firms to decide when they want to embrace the changes to their business model that a meaningful investment would necessitate. For the moment, the industry is at the beginning of the curve, so lawyers have the choice to become early adopters, or to wait until industry adoption becomes widespread before committing.