AI and the end of lawyers: Myth or reality?

Jean-Gabriel Ganascia defines Artificial Intelligence (“AI”) as “the discipline in computer science that seeks to develop machines that mimic the superior faculties of intelligence”^[1]. In practice, when applied to the legal industry, AI could be used for a number of tasks, including, but not limited to, the generation of legal documentation, the definition of strategies in disputes and litigations, the automation of basic tasks commonly performed by trainees and junior associates. In this context, should we fear the progressive end of lawyers?

While legal technologies are developing at high speed, many people wonder about the need – if not the benefits – to call upon a lawyer to draft legal documentation (1), or to manage litigation (2). Furthermore, as some basic tasks performed by lawyers are easily automatable (3), should we fear a progressive end of the legal profession (4)?

1. Drafting of legal documentation

For the drafting of routine legal documents, such as the by-laws of a single-member start-up company, AI can easily do the job. When there are no several shareholders/partners, the legal risk is indeed moderate.

As such, there are today many websites that, based on a questionnaire completed by the user, generate all of the legal documentation required to create a company, and provide assistance for the completion of the incorporation formalities with the clerk of the competent court.

On the other hand, for complex documents – such as, to continue with start-up companies, the shareholders’ agreement that must be signed wherever funds are raised[2] – the assistance of a lawyer can be helpful in many aspects:

• Firstly, because lawyers will deliver a customized advice and, to protect the founder, put in place devices adapted to his/her profile, needs and expectations (does the founder plan to gradually withdraw from the project or does he/she wish to keep control?), to the profile of the third-party investor (is the investor a finance professional, a relative, etc.?), to the terms of the transaction (acquisition of a shareholding interest or debt financing?), etc. As a matter of fact, each situation is different and requires a tailor-made strategy.
• Then, because lawyers, contrary to expert software systems, must comply with a set of ethical rules, foremost among which are confidentiality of communications and professional secrecy. Yet, can anyone today guarantee the confidentiality of the data exchanged between the user and the machine/software system?
• Lastly, because lawyers, contrary to the websites referred to above, must take out a professional civil insurance policy (also known as malpractice insurance) that covers the mistakes that they may make in delivering professional legal services.

2. Management of litigation

In litigation matters, predictive justice, using case law collections, now makes it possible to:

• predict the chance of success of a case,
• assess the amount of damages that may be awarded,
• suggest avenues for strategy.

All of it, on the basis of the decisions rendered by the relevant Court.

However, here again, AI has its limits: Predicting the outcome of a case is conceivable and rather easy wherever the Cour de cassation (French Supreme Court) has issued several consistent rulings on the same legal question. But how could, for instance, a software system anticipate a potential case-law reversal that a lawyer could foresee given, in particular, the evolution of applicable legislation or a change in mentality?

3. Automatable basic tasks

These tasks are primarily legal research tasks. An increasing number of tools are being developed in the hope of delegating entirely this type of works to machines. One of the best known is ROSS (a product developed by IBM) that can, based on data communicated orally just like during a standard discussion between colleagues, conduct research on a specific point of law in US bankruptcy laws, and then deliver a rough draft of memorandum.

The legal industry is not the only industry affected by this phenomenon: Many surveys show that in every sector the task of the youngest may be impacted insofar as they are automatable.

4. What future for lawyers?

The question of how AI will impact the world of work unleashes emotional debates: Some have issued very alarmist opinions, suggesting that the emergence of IA will destroy a considerable number of jobs in the years to come. Others are more tempered and anticipate instead an evolution of our ways of working, through the adaptation and the re-definition of certain professions.

To refute any cataclysmic vision, here are a few figures to measure the actual impact of AI: According to an analysis made by McKinsey & Co in 2015, fewer than 5 percent of occupations can be entirely automated and about 60 percent of occupations could have 30 percent or more of their constituent activities automated[3]. Regarding more specifically the legal profession, according to a more recent study carried out by two eminent US Professors, only 13 percent of the legal works done by a lawyer could be automated[4].

[1] Jean-Gabriel Ganascia, Intelligence artificielle : vers une domination programmée ?, Le Cavalier Bleu éditions, 2007.

[2] Injection of funds by a third-party (e.g. an investment fund, an investor or even a relative) to contribute to the growth of the start-up.

[3] Four fundamentals of workplace automation, McKinsey Quarterly, 2015.

[4] Can Robots Be Lawyers? Computers, Lawyers, and the Practice of Law, Dana Remus (University of North Carolina School of Law), Frank S. Levy (Massachusetts Institute of Technology (MIT) – Department of Urban Studies & Planning), 2016.