Paper delivered on the opening day of the XVII National Assembly of the Observatories on Civil Justice Bologna, 13-15 June 2025.
[The original version in Italian is published here: La formazione dei giuristi ai tempi dell’intelligenza artificiale]
Abstract
From 13 to 15 June, the National Assembly of Civil Justice Observatories took place in Bologna. Luciana Breggia invited me to discuss legal education in the age of artificial intelligence, a topic close to my heart, and the potential impact of artificial intelligence on the work of lawyers and judges.
I want to highlight three points.
- Firstly, the study of law should focus on the dynamic interaction of formants. If it deals only with one formant, however, it does not study law for what it is, but sees only a fragment — an artificial cross-section, or even a partial one.
- Studying law also means ‘putting it to the test’: comparing it with real and hypothetical cases, imagining and evaluating its application to reality, and understanding its implications. This is very different from merely receiving and regurgitating content.
- The truly disruptive impact of AI will be to enable us to achieve things that we are not currently capable of. AI could enable extraordinary advances in all fields. However, it could leave many people behind who are unable to take advantage of it. AI could be an agent of rebalancing opportunities or a multiplier of critical issues, such as laziness, prejudice, inequality and discrimination.
Recently I have had several occasions to speak about legal education in connection with the use of artificial intelligence.
In this talk I would like to broaden the perspective: I will start from what legal knowledge ought to encompass, move on to how legal education should be re-thought, and finish by sharing a few ideas on how legal knowledge can be enhanced by artificial intelligence.
Law in Action and the Dynamics of Formants
I remember attending my very first law school lecture: the lecture theatre was packed, and it was just one of several parallel courses attended by around 5,000 new students. The topic was ‘sources of law’, meaning ‘statutes, regulations and customs’, as listed in Article 1 of the Preliminary Provisions to the Civil Code.
Only after a few years, in a small comparative-law course of about thirty students, I encountered eye-opening works by masters such as Gino Gorla, Rodolfo Sacco and René David.
One discovery was Sacco’s theory of formants, a model for understanding the law in action, meaning the dynamic interaction of written law, case-law and scholarship. These elements, which Sacco calls formants (borrowing the term from linguistics, also because “sources” already had another meaning), I believe that today should be added: supranational (especially EU) legislation, uniform law, European case-law, model contracts (as leasing, factoring, independent guarantees), decisions of independent authorities and what is called soft law (guidelines, codes of conduct and ethics, circulars, notarial recommendations, protocols, etc.).
In this interaction, it would be reductive for scholarship to limit itself to observing and interpreting only one of the formants (written law, case law, academic studies by others, and so on). The task of scholarship should first of all be to recompose the system, inspired in particular by constitutional principles and, above all, by the fundamental principles of equality and the centrality of the person (Articles 3 and 2 of the Italian Constitution). In doing so, scholarship should examine and anticipate the dynamics and trends of the formants, their points of balance and their stability or instability.
Sacco taught us that law is the result of the dynamic interaction of its formants. When these formants are not convergent, the equilibrium becomes unstable and will change sooner or later. Consider the concept of non-compensability of damage to legitimate interests, for example. My first significant case as a trainee involved seeking compensation for exclusion from a public tender that had already been deemed unlawful by the administrative court. At the time, case law uniformly denied compensation for violations of legitimate interests (i.e. unlawful exercise of administrative discretion). However, Article 2043 of the Civil Code clearly states that all damage caused by an unlawful act is compensable, both then and now. Innovative scholarship had already pointed out that such case-law lacked solid foundations in statutory law and even less in constitutional values. Consequently, there was a discrepancy between case law and the written law, as well as part of the scholarship.
We represented a company that, because of that unlawful exclusion, had been forced to cease its activity. Led by my dominus, I attempted to present the case in terms of new categories that were emerging in other areas of law, such as loss of chance doctrine. I was hoping to breach the consolidated jurisprudence by making a subtle distinction. However, the court dismissed the case in just two lines: ‘This concerns a legitimate interest; therefore, it is not compensable’. The client gave up appealing. Unfortunately for them, the landmark Supreme Court judgment No. 500/1999, delivered shortly afterwards, with a single stroke equated legitimate interests and subjective rights for the purposes of compensation under Article 2043 of the Civil Code. Case law aligned with written law and less rigid scholarship, establishing a new equilibrium that has never been questioned since. Therefore, the study of law should focus on the dynamic interaction of formants. If it considers only one formant, it does not study law as it is; it sees only a fragment — an artificial cross-section that is incomplete from the outset.
Teaching Law
Studies on adult education show that adults are only open to learning if they are motivated, and that this motivation primarily arises from experience. If you try lecturing an adult on something they “must” know but are not interested in, after two minutes their mind will wander, they will reach for their phone or close their eyes.
This phenomenon has its own discipline of study, andragogy, which is derived from pedagogy but focuses on adults. The idea is that adults cannot simply be filled with content like vessels. However, this concept does not only apply to adults. To varying degrees, children and young people also learn best — indeed, truly learn — through experience (learning by doing), as pedagogues such as John Dewey, Maria Montessori, Don Lorenzo Milani and Tsunesaburō Makiguchi have demonstrated.
Try delivering a monotonous ex-cathedra lecture to teenagers and see if they pay attention. The level of distraction increases the more they are exposed to other learning stimuli, whether positive or negative. I believe that today’s youngsters do not listen much, not because they are lazy or unmotivated, but because they are already ‘adults’ in a sense and require motivation to pay attention. They experience flows of information — broadly, of ‘learning’ — that are incomparably richer than those we experienced at their age. They cannot see themselves as empty vessels; they need to perceive the utility of learning. They have access to all the answers, both right and wrong, at their fingertips; their problem is to distinguish the ‘right’ ones. More precisely, they need to learn how to select answers and, before that, how to formulate appropriate questions. In a word, they need to develop critical thinking. For law students this should enable them to study law as scholars ought to: composing the formants into a system and deriving operative rules. All the more so if they aim to become judges, lawyers, notaries, officials or senior civil servants.
Studying law therefore also means “testing” it: confronting it with real and hypothetical cases, imagining and assessing its application to reality and grasping its implications — something very different from the mere reception and bland repetition of content.
Artificial Intelligence
Finally, what can we do with artificial intelligence?
I often hear that AI will free us from the simplest and most repetitive tasks. This is true, but to limit AI’s contribution to such tasks would be reductive. Yes, it will help in this respect, but that is not the most important aspect.
There is also a risk of encouraging laziness: we may achieve a goal without directing ourselves or evaluating the outcome’s meaning. However, laziness and superficiality did not emerge with AI; in fact, AI could highlight and address these issues.
I was the only person who used a computer. I had only started using one a year earlier, to write my degree thesis. Pleadings and opinions were initially handwritten before being typed by the secretary on a typewriter. Corrections were made with a pen, an eraser or correction fluid. Passages from earlier documents or judgements were copied and pasted using scissors and glue.
Although computing made that kind of secretarial work obsolete, it also enabled things that were previously impossible. Digitalisation allowed us to access databases, first via CD-ROM and then online. Rather than waiting two years for them to appear in printed repertories, we can now consult precedents in real time. The internet has given us email, social networks, digital news, Netflix and Spotify — all of which did not exist before. Now, AI has opened a new chapter thanks to its rapid development and disruptive potential.
So far, digitalisation has supported legal professionals behind the scenes in their preparation, and then it has transformed infrastructure through e-filing systems. AI will achieve much more in these areas. However, its most significant impact will be enabling us to perform tasks that are currently impossible.
Consider pleadings and legal opinions. Although email has made it possible — and mandatory — to deliver opinions almost in real time, their preparation often follows methods that have remained unchanged for centuries. Pleadings have not changed either, except that they are now typed on a computer. These methods are inefficient and often ineffective. The only genuine directive has been the requirement for ‘clarity and conciseness’, which was previously developed by the Observatories on Civil Justice. Other reforms have mostly concerned page or line limits.
To illustrate how the work of lawyers and judges may change in the future, I often cite the example of a Milan law firm that hired a design company to create an interactive brief to help a judge navigate through thousands of pages of evidence. This was undoubtedly a more effective way of communicating the contents of the file, but a few years ago it required months of work. With AI it could be done in a few hours, providing a more efficient working method that does not currently exist or would be too costly.
I imagine the jurist of the near future, equipped with a dashboard of AI tools and interacting with powerful instruments never before seen. Returning to the formants paradigm, we will be able to analyse formants in their dynamic interactions and compare judicial and scholarly trends through the lens of statutory law, and vice versa. All this information will be readily available on one’s dashboard. AI is a gateway to something akin to superpowers: it won’t make us work less, but it will enable us to produce more and better.
Before closing, let me recall Spiderman: with great power comes great responsibility.
I like AI, especially interactive AI. I have always been better at asking questions than providing correct answers, so AI is the perfect complement — a tool for practising modern maieutics, or prompting: the art of asking and answering the right questions.
However, despite my enthusiasm for AI, I am equally concerned. While AI could bring extraordinary advances in every field, it could also leave many people unable to exploit it behind. AI could redress imbalances in opportunity or exacerbate existing problems such as laziness, bias, inequality and discrimination.
Rules are therefore necessary. I am convinced that rules are not a restraint, but rather a prerequisite for technology to be useful and socially and politically acceptable to people.
In my view, the overriding priority today — the foremost responsibility of governments, businesses and individuals — is to undertake a massive collective training effort to educate people in the use of AI, and to raise awareness of its risks and impacts. It is no coincidence that the first provision of the AI Act to come into effect on 2 February 2025 is the obligation for all employers to train their staff. This is a significant commitment, given the disruptive impact of this technology on our lives, for better or worse. However, although the law exists, awareness of its importance does not. This is evident in governments, who possibly prefer to accumulate laws rather than design operational plans, and in companies, who often view AI solely as a means of reducing production costs rather than as a tool for redesigning their services.
In any case, one thing is certain: nothing will stay the same. We are in the midst of a movement of perhaps unprecedented magnitude, driven by its own self-reinforcement and characterised not only by its scale, but above all by its speed and acceleration. I am sure that we will have many new topics to discuss at next year’s annual Observatories on Civil Justice meeting.
Agostino Clemente
(Lawyer, Observatory on Civil Justice of Rome)