Artificial Intelligence and Law
MONGOUE YANGOP FRANCIS DERRICK, EO-204i, KNEU
What is palpable today is the confluence of the legal world, which is an environment of tradition and ancestral values evolving thanks to new technologies. These technologies are not necessarily disruptive, they are sometimes reasoned. Artificial Intelligence can seize the right.
“Code is Law”, the superposition of this famous Lessig maxim, provides an interesting insight into the politics of given architectures. The A.I. will certainly bring about many upheavals, but lawyers will know how to deal with it. The challenge will be to ensure that it is AI that is invested by lawyers and not vice versa.
The rise of Legaltech, these famous law startups, seems to have taken the lead in terms of digitalization of the legal sector. The use of technology and software to offer legal services and make the law more accessible, implies that the Legaltech sector disrupts the law in its own way, very often based on developed and efficient algorithms.
The example of Machine Learning shows how Artificial Intelligence can push an ancestral matter, such as Law, to its limits.
Machine learning is the very definition of learning artificial intelligence that improves its performance in solving tasks without being explicitly programmed. There is no specific legal framework in this area and everything will depend on the purposes provided for the algorithms implemented.
Besides, in order to function, Machine Learning algorithms are based on training sets and sometimes require personal data to function. This therefore entails reinforced documentary obligations, in particular with the entry into force of the GDPR. Indeed, the controller must, by himself, implement all the measures necessary to comply with the law in force.
It will be a question of seeing to what extent the law can be seized by artificial intelligence, that is to say, to what extent artificial intelligence will have a real and effective influence on the legal field. To do this, it is necessary to study the impact of this intelligence on the legal professions, on the functioning of justice or even try to answer the questions of transparency and ethics raised by the development of these new technologies.
IMPACT OF ARTIFICIAL INTELLIGENCE ON LAW JOBS
The growing development of artificial intelligence and its impact on the legal professions raises questions about the place and role of the lawyer of tomorrow. The way of doing Law will evolve in the coming years and in reality has already evolved a lot.
The automation of internal processes, the development of smart contracts, increasingly intelligent algorithms, are all issues that the legal professions must face and anticipate as much as possible.
How? ‘Or’ What?
The rise on the legal market of Legaltech companies, some of which have developed digitalization solutions that are already very advanced in certain sectors, making it possible to provide effective responses to the challenges of digitalization, particularly for a legal department.
Two schools exist; a first, which we will voluntarily call that of “immobility”, advocating the status quo, opposed to changes and promoting the fact that artificial intelligence is there to replace the lawyer in the long term, thus opposing the development of technology and the traditionalist practice of Law. This first movement increases reluctance and mistrust vis-à-vis digital.
Another school, which we will call “progressive”, is favorable to the development of a mastered technology which will make the legal professional an “enhanced lawyer”. This second school will allow effective collaboration between digital and law, allowing the lawyer to seize these new digital tools, thus facilitating his daily work by freeing himself from repetitive and redundant tasks.
It is certain that AI will be able to provide great services to the world of law, in particular in an effort to rationalize legal information, thereby facilitating research.
The limits of artificial intelligence lie in the fact that legal reasoning remains specific to humans. The construction of a syllogism around premises seems to belong, at least for the moment, exclusively to the lawyer. Adaptation, interpretation, emotional, are all attributes that the machine does not have and it is in this that the paradox appears; on this point, would it not rather be artificial intelligence which would be seized by the Law?
Thus, for the profession of the lawyer of tomorrow, artificial intelligence should be seen as a tool at the service of the lawyer, certainly helping the latter to become more efficient in responding to more complex requests. The legal professional will then focus on tasks with higher added value.
In general, the legal profession in the broadest sense of the term remains a sector in which there is a high added value in terms of the work carried out.
Hence, for the legal professions, the impact of this growing technology must be put into perspective; human intelligence remaining the keystone of the legal structure.
FUNCTIONING OF JUSTICE
“Predictive” justice is a concrete example of how artificial intelligence grasped the law. Raising serious concerns that the judgment would be rendered on criteria based on what judges have already judged. Legal limits already exist and technical solutions are already proposed, in particular with open source code or verification tests from data sets.
Reluctance among judges is greatly felt, seeing the traditional French system of Civil Law, switch with this phenomenon to a Common Law system based on the previous one.
The establishment of such justice would call into question the classic model on which our Law is based. Indeed, predicting court decisions on the basis of decisions already rendered, is therefore to progressively move from a Civil Law system to a Common Law system.
Although it is obvious that predictive justice makes it possible to better anticipate and quantify the chances of winning or losing a trial, it should not, ultimately, be the technological solution to replace the judge.
However, this prediction willingly participates in this process of relieving congestion in the courts, because it dissuades certain litigants from initiating proceedings based on their chances of winning or not winning a case. It will certainly favor modes of amicable settlement of conflicts by favoring mediation, transaction or even conciliation.
Predictive justice therefore pushes for the transaction and can reduce the clutter of the courts, faced with budgetary underfunding. It is in fact the law of 2016 for a digital republic which brought into our law the making available to the public free of charge of all anonymized decisions.
Thus, this law paved the way for predictive justice. In this area, artificial intelligence is seen as the “new fantasy” of objectivity for justice. But once again, it would be better here to observe AI as an assistant, a facilitator and not as a replacement.
To assist the judge or the jurist in his decision-making is to bet on an artificial intelligence which collaborates with the human being.
Thus, the path to fairer justice would pass through the path of artificial intelligence.
With its empowerment and the completeness of its big data, artificial intelligence would make the delivery of justice certainly more reliable. The jurist would no longer be considered as the exclusive holder of the faculty of appreciation and judgment of a legal situation but would be supplanted, if one follows what one said previously, by a new instance of truth estimated fairer and more viable.
However, let’s not forget that it is humans who are at the origin of the work created by algorithms, they must be constantly evolving and adapting. Human remains at the heart of the system, the question is more about who will do the law for tomorrow, whether the lawyer or the engineer will program the algorithms to make them work. It is in this sense that artificial intelligence will seize the Law.
The main risk linked to the implementation of this system would be that artificial judicial intelligences will eventually be able to go beyond the framework of a single predictive justice to turn to a single justice rendered by the machine. This risk will be felt in particular with the lack of a reversal of case law, or of thorough and specific legal analyzes on a given point of law.
If the learning artificial intelligence (or strong) was able to make a court decision, it is different with regard to transparency regarding the choices made by the machine. If the faults of the machine are very often the faults of the men who programmed them, it will be necessary to understand why this decision was taken in one direction and not in another.
If we can no longer understand how artificial intelligence made a decision, it is that the Law would actually be at the service of this intelligence and that the latter has largely grasped the legal matter
THE CHALLENGE OF EXPLICABILITY
An algorithmic decision is said to be “explainable” if it is possible to easily account for it from known data and characteristics of the situation. Explaining how an artificial intelligence arrived at a result, would be the prerequisite to try to avoid that it effectively seizes the Law by having the impossibility of explaining its reasoning.
Indeed, the AI would effectively seize the Law if it was impossible to understand how the machine made this or that decision.
The entry into application of the GDPR, the strengthening of existing legislation, the law for a Digital Republic indicate the desire for transparency and loyalty in the use of data by algorithms.
Transparency in the execution of algorithmic processing is a major issue because it requires a better understanding of the decision-making process of the algorithm. In order for individuals to place their trust in artificial intelligence, it is necessary to establish the highest transparency and loyalty in the algorithmic processing that drives it.
Behind the notion of explicability hides the notion of “mastery”. The mastery of the lawyer in the development of algorithms of a learning artificial intelligence. Indeed, if there is mastery, the phenomenon of capture of the Law by AI will arise much less.
However, providing more and more data to artificial intelligence promotes its development and contributes to the idea that it becomes less controllable and more autonomous. This will imply that AI will gradually take control of legal matters by, in particular, solving repetitive legal tasks with low added value.
Explainable AI is also desirable due to the fact that it will improve performance and help decision-making. Technology would then be thought of as a catalyst to increase human decision-making.
Artificial intelligence is seen as a technological attempt to elevate computer systems to cognitive capacities of knowledge, reasoning and learning comparable to those of man although they are considered to be his own.
Thus, artificial intelligence will certainly bring about many upheavals, but lawyers have been able to deal with several of them in the past. Artificial intelligence is therefore a field for lawyers to invest in, not the other way around. The paradigm must be reversed, the law can seize artificial intelligence.
When asked if artificial intelligence challenges our way of doing law or does it offer new perspectives for the professions of law, you will understand that we will have to evolve towards an optimistic, collaborative and daring future.