The term „Machine Learning“ was unheard of in most law firms years, or even months ago. Now it seems almost every discussion around the „digital law firm“ of legal tech resolves around this topic. Simply throwing the term “Machine Learning ” into a discussion guarantees attention and excitement. The media continuously reports that Machine Learning will either destroy the legal profession or lift it up to new heights. Comparatively, there has been very little discussion on what this developments actually means for law firms and, perhaps most importantly, how they could get Machine Learning ready.
The term „Machine Learning“ is also often used interchangeably with Artificial Intelligence. Thus before we dive into the question of how law firms could get Machine Learning ready I want to briefly highlight the differences between Machine Learning and Artificial Intelligence.
- Artificial Intelligence is the broader concept of machines being able to carry out tasks in a “smart” or „intelligent“ way.
- Machine Learning is a current application of AI that „gives computers the ability to learn without being explicitly programmed“ (Arthur Samuel, 1959). Machine learning is closely related to (and often overlaps with) computational statistics, which also focuses in prediction-making through the use of computers.
The use of Machine Learning is already changing business in virtually every industry. Even though the likelihood of self-directed, quasi-intelligent computational robots emerging in the foreseeable future is extremely low, Machine Learning can provide huge benefits in the day-to-day business. Over the next 5 to 10 years, the biggest business gains will stem from the gathering and sourcing of data and providing this data in new products and/or business models to the customers/client. Machine Learning will accelerate and facilitate finding patterns and automate value extraction in many areas.
So how can a law firm start to use Machine Learning?
# 1 Step: Analyze business processes
Machine Learning is perfectly suitable for repetitive processes. Therefore, the first step is to analyze processes in a law firm in order to detect tasks, procedures and decisions that are repeated frequently and consistently. This could be, for instance, various tasks in a Due Diligence (DD) process. A DD process comprises of every task, procedure or decision that is done, followed or made from the initial contact with the client and the delivery of the DD report. After having identified these processes it is vital to gather as much data as feasible around them. This is the kind of data that can be used to fuel Machine Learning in the future.
# 2 Step: Understand Machine Learning
Machine Learning won’t be suitable for all processes. It is not a solution for every type of problem. There are a lot of problems and areas where robust solutions may be developed without using Machine Learning techniques. Machine Learning is, for instance, not required if you can determine a target value by using simple rules, computations, or predetermined steps that can be programmed without needing any data-driven learning. Often these robust solution suffice to gain transparency and reliability.
Machine Learning is useful if the set of rules is unclear or follows complex, non-linear patterns. In particular, Machine Learning is helpful where
- Rule-based solution are inadequate: If a large and overlapping number of factors influences the answer it is impossible to use a simple (deterministic), rule-based solution. It is just too complex to accurately code the rules and to fine tune them. Machine Learning, however, effectively solve this problem; or
- Large-scale problems arise: ML solutions are effective at handling large-scale problems. A smaller set of data (documents etc.) might be handled manually or with rule-based solution. This however becomes impossible for millions of documents.
# 3 Step: Focus on quick wins
Focus on the low hanging fruit. Automation based on Machine Learning will work best on well defined and understood processes which are not overly complex. In a legal context, focus on tasks, decisions and problems that do not require a lot of in-depth „legal thinking“. Concentrate on tasks that are repetitive and could be performed by literally „everyone“. Surprisingly, there are lot of tasks in a law firm that qualify for Machine Learning testing ground, i.e …
#4 Step: Action