On one hand, Machine Learning has been developing for the last couple of decades, and during the past years has been on the spotlight for several industries. On the other hand, legal profession is one of the most ancient fields in the world. So, how can we relate machine learning to the application of the law in these days?

About being a lawyer

Many people may think that being a lawyer is like being Harvey Specter or John Milton: Closing big deals, wearing glamorous suits in fundraisers and dazzling jurors with emotive speeches about reasonable doubt and guiltiness. That may be true for 1% of the lawyers around the globe, but most legal work includes long research hours, gathering and analyzing data for clients, and comparing probabilities when preparing for a case.


So, even if some lawyers obtain the spotlight in big cases matters, there is a team of hard-working paralegals and junior associates that work 14-16 hours a day to get the proper data to their superiors in order to prepare for big litigation cases or closing deals (every lawyer has gone through that process).

What can Machine Learning do for the lawyers?

Picture the scenario: 11:00 p.m. in the office and a Junior Partner comes to your paralegal cubicle and asks for a thorough research on jurisprudence and court precedents for - you name it -: discrimination of gender, merger control in the pharmaceutical business or [insert long and tedious topic here]. Of course, the deadline is 09:00 a.m. the next morning. The typical form of doing this is looking for a keyword of the case and then start reading all this (long) related cases until you figure, hopefully, a track of action before 09:00 am in the morning. Well, apparently, there is another way! A group of young guys graduated from a Stanford Law with a digital and more technological approach has developed a tool to make this job easier: Ravel Law. By using data visualization, analytics and machine learning, Ravel Law enables lawyers to find, contextualize, and interpret information that turns legal data into legal insights.

By this type of tool, based in machine learning, lawyers can save several hours preparing for cases and doing research. Ravel’s data base is huge and contains information of more than 5 decades of cases. What’s more, the filters that can be applied are really varied. As an example, two of the functions that this tool provides are (i) "Relevance", which aligns cases vertically, with the most relevant cases moving to the top of the screen, while (ii) “Ravel” organizes cases with a gravity model, so cases that are heavily cited by other cases in your search results will be the center of gravity and pull other cases towards them.

The use of this type of services definitely saves Firms a lot of (usually) non-billable hours, which can be addressed then working in another cases. Time is of the essence, especially when you work in a Law Firm, and machine learning is making the time management more effective for lawyers. As we can see, machine learning and the law are getting closer, and it should not surprise lawyers to have more and more of these tools later on in the future.

What does Suits have to do with this?

Not a lot actually. Besides that I really like the show, I just wanted to point out that seeing the likes of Harvey Specter in the big and small screen might be misleading to people not related to law practice; and thus, the reality is that not everybody that goes to Law School, can automatically become Harvey Specter (especially without researching (a lot!) first).

Picture: The black tie blog
Machine Learning
Machine learning is based on algorithms that can learn from data without relying on rules-based programming. It came into its own as a scientific discipline in the late 1990s as steady advances in digitisation and cheap computing power enabled data scientists to stop building finished models and instead train computers to do so. The unmanageable volume and complexity of the big data that the world is now swimming in have increased the potential of machine learning—and the need for it.

People learn from experience - and now machines do so, too.

Back in 1967, McKinsey published an article by Peter Drucker named "The manager and the moron". In it, it states that "the computer makes no decisions; it only carries orders. It's a total moron, and therein lies its strength. It forces us to think, to set the criteria. The stupider the tool, the brighter the master has to be - and this is the dumbest tool we have ever had".

Half a century later, this statement couldn't seem further away from the truth. Progress has been striking: computers now can process human speech and answer questions, make recommendations and performs a series of actions. They can even recognise objects and patterns in images. Computers are replacing skilled workers in a large variety of fields - ranging from medicine and architecture to aviation and the law. In fact, a recent study carried out by Oxford University concluded that over 47% of all U.S. jobs are susceptible to computerisation!

So, should we be racing with or against intelligent machines?
While the thought of rebellious robots taking over our society has been the subject of many science-fiction books and movies for decades, digital innovation has only started to scratch the surface of what we can truly accomplish through technology. At Machine Learning Hub, we will explore both sides of the AI controversy, analyse current trends and business applications of machine learning (as well as its subfield of deep learning) and examine future developments in the field.

Image source: Uni Hamburg