How to replace lawyers with legal artificial intelligence

How to replace lawyers with legal artificial intelligence

Can you tell the jury how legal artificial intelligence software made lawyer profession extinct? I admit I was looking for a catchy title when I wrote this. Today we are far away from replacing lawyers with algorithms, but learning machines are seriously beginning to take over some pieces of their job, so we have to recognize there are forces able to reshape the legal industry. And will develop faster than you expect.

According to common experience, a lawyer spends the majority (40%) of his time dealing with documents and making researches. The rest is about writing (30%) and communicating (20%), while a minor portion goes to travelling. A machine would still find extremely difficult to interact with a client or a court, but can search for documents pretty well and there are many progresses about software writing standard stuff.

Once again, WATSON, the IBM cognitive computer is guilty. A group of developers built on its technology Ross, a huge database of thousands of legal documents. Ross can acquire unstructured data, in the form of text (not necessarily located in the rows and columns of a standard database), and answer to a question posed in natural language. In other words it works like Siri, but it’s specialized on a couple of legal practices. It can analyze huge volumes of historical information in seconds, is never tired and very fast. Every answer is associated with detailed references ordered by importance. And continues to learn, because a human can approve or reject the answer, asking for integration.

How to replace lawyers with legal artificial intelligenceAlso when it comes to writing, legal artificial intelligence algorithms can help to face some kind of issues. In particular they can try removing some kind of ambiguities. Legal wording is written in natural language, so the scope of connectives such as “and” and “or” can be unclear, the use of “unless” is also capable of several interpretations, and for example “if and only if” can be different than a simple “if” etc.… This activity is included in the wider area of research called “Legal Reasoning” and researchers are trying to develop formal models to address the challenge. In general terms, models explore how legal experts analyze problems, create arguments, and make decisions. It’s about a century that different theories have been proposed: in 1913 Wesley Hohfeld, a legal positivist, suggested that all legal states of affairs can be described in terms of four fundamental relations: rights, duties, powers and liabilities; but in 1948 Jerome Frank argued that law cannot be characterized as a formal system, because the infinite variety of particular facts on which legal decisions turn make legal reasoning inherently unpredictable. I’m not a lawyer and even if I’m instinctively a realist rather than a positivist, machine learning and artificial intelligence are slowly proving law practice can be codified. A very good paper written by Carole D. Hafner at the College of Computer Science, Northeastern University, Boston can be downloaded here.

When legal artificial intelligence algorithms face quite structured and standardized pieces of legal activities, they can support lawyers. But there’s something more than a software aid. Artificial intelligence analytics is entering in the area of strategic guidance. In fact, it can crawl through records of past cases, analyze similarities and differences with a specific case and produce models able to predict the likelihood of cases being won or lost. It was at the end of 2014 when London firm Hodge Jones & Allen (HJA) teamed up with a leading academic to pioneer a predictive model of case outcomes that will enable the firm to better assess the viability of its personal injury caseload. That was a successful interaction between statistical techniques (to assess the probability) and artificial intelligence (to identify the right items and patterns to analyze).

Please share this post, if you don’t mind and if you like it

How legal artificial intelligence is going to reshape the industry

So back to the provocative title of this post, I feel we don’t have enough elements to say that a machine can replace a lawyer today or that it will happen in the next 5 or 10 years. Two researchers, Levy and Remus, broke down legal tasks into pieces that are likely to be strongly, moderately, or lightly affected by AI: document review, for example, is easy to automate, but it is hard to do the same with negotiation. They concluded that only about 13 per cent of legal work will be taken over by computers within the next five years. Abstract & Paper.

Ironically, the same source, publishes a paper titled “The Great Disruption: How Machine Intelligence Will Transform the Role of Lawyers in the Delivery of Legal Services”, which describes five areas in which machine intelligence will provide services or factors of production currently provided by lawyers: discovery, legal search, document generation, brief generation, and prediction of case outcomes. Abstract & Paper.

The replacement of humans with machines, as usual is not imminent and probably will never affect 100% of a profession, but I see enough hints that legal artificial intelligence is going to change this industry.

How to replace lawyers with legal artificial intelligence


First, the path of growth of young lawyers will completely change. Juniors today spend their time exactly searching for documents and examining past cases. If a machine will do this on their behalf, legal firms won’t need them anymore and / or they will have to grow their knowledge in other ways. In an extreme scenario, where legal firms adopt legal artificial intelligence algorithms which are more accurate and less expensive than a human, young inexperienced lawyers simply won’t be hired. At some point, after a couple of generations, experienced lawyers won’t exist anymore. Optimists simply say that lawyers will evolve their skills becoming more technological competent. I’m quite skeptical. I can’t see a lawyer coding to understand if the machine is returning the right results. That’s another job and anyway humans are not “programmed” to control machines.

The second aspect I see is competition. Startup young legal firms can enter in the market and cut the price of their services, because a platform can deliver the right answers straight away. The assistance to routine cases, those without a great complexity, will become commoditized; probably LAIS (legal artificial intelligence as a service) will pop up. New entrants will be looking for a competitive edge, have great freedom to design processes from scratch and probably access to external sources of capital, as they will be perceived as innovative startups rather than old economy big legal firms. On top of it, other industries might adopt the same technology and step into the legal industry. Auditors, accountants and several other finance professions might be wiped away by automation, so they will be tempted to use artificial intelligence to enter the legal arena. Industry borders will become less defined.

The last item is the public reaction to all of this. A large portion of the society will be definitely happy that the cost of legal assistance decreases and access to justice becomes more democratic. Another group won’t accept to be judged or supported by a machine; just because an algorithm might perform efficiently it does not mean that it is correct. A minority will push artificial intelligence to the limit for their purposes: why not tailoring a research or a defense strategy to the Judge who presides over your case or using the algorithms to find weaknesses in the regulations or incongruities to circumvent the laws.


Newsletter: because there’s more than legal artificial intelligence here!

The Futurist Hub Newsletter is the greatest thing after the Big Bang. Once per month, only the news, free of spam. And with a free ebook as a bonus.

Prev Why it makes sense to compare human and artificial intelligence
Next Enzo Ferrari – An Inspiration for tomorrow


No Comments Yet!

You can be first to comment this post!

Leave a Comment