Scholarship

As the regulatory landscape grows increasingly complex, legal analytics is emerging as a transformative tool for risk management and compliance. By applying advanced data analysis techniques such as machine learning and natural language processing to vast volumes of legal and regulatory data, firms can shift from static, retrospective approaches to dynamic, forward-looking strategies. Legal analytics enables organizations to identify trends, predict risks, and make informed decisions, fostering a more proactive and strategic legal management framework.

This chapter explores the evolving role of legal analytics in business, highlighting its applications in privacy compliance, fraud prevention, and legal counsel monitoring. By contextualizing legal analytics within the broader discipline of Law & Management, this chapter underscores its potential to align compliance with competitiveness, redefine legal risk, and bridge the gap between legal and strategic decision-making in today’s data-driven business environment.

Large Language Models (LLMs) have crashed into the scene in late 2022, with ChatGPT in particular bringing to the mainstream what has before this remained within the domain of the initiates. This paper introduces the main features of LLMs and related Artificial Intelligence (AI) to the legal community, while reviewing their potential application in a legal context, as well as the main questions and issues raised by their increasing presence in a jurist’s life. Adopting a structural approach, the analysis highlights the areas of legal activity that stand to gain – or lose – from the generalisation of LLMs in our workflow. The radical innovation represented by LLMs will force jurists to rethink their approach to the law, their own role in it, and the future of legal education and training.

S’il en faut croire la plupart des commentateurs, la qualité du droit français est en baisse perpétuelle. Remédier à cette dégradation est l’objectif des acteurs publics depuis plusieurs décennies, de circulaires en guides de légistique. Ces tentatives ont-elles cependant permis une amélioration de la « qualité » de la loi ? A l’aide des données obtenues sur Légifrance et des outils de « natural language processing », et des recommandations légistiques, il est possible de retracer en partie les évolutions qui ont été ou non réalisées dans la rédaction du droit français. L’analyse met en lumière des améliorations probables de forme, mais la qualité des lois nouvelles demeure une question qui échappe à la résolution.

"Reading" is the primary technology investment arbitration practitioners use to engage with party submissions, prior awards, investment treaties and all other written materials relevant to the arbitration. Yet, an alternative and complementary approach to parse investor-state arbitration (ISA) texts is rapidly emerging. In what is variably known as "text mining", "text analytics", or "computational analysis", algorithms rather than humans are beginning to digest written materials to extract relevant insights. What this algorithmic approach lacks in nuance (computers are bad at understanding text), it makes up for in efficiency (computers excel at crunching numbers). This chapter targeted at arbitration professionals introduces the methodologies underpinning text analytics, explains and evaluates their use cases in investment arbitration, and assesses their promises and limitations.

International litigation and arbitration hinges on the use (and, sometimes, misuse) of “authorities” (i.e., precedents and the teaching and writings of “publicists”). Parties spend time and effort marshalling supportive authorities, distinguishing or refuting unhelpful ones, and monitoring legal developments in unrelated cases and in academic debates to bolster their legal arguments.

This practice has profound implications, for many key issues of international law (sources, role of precedent, etc.) are also questions about the nature, scope and relevance of authorities in international dispute settlement. There is however little research on the practice of citing authorities in international disputes, and even less empirical research that sheds light on such citations.

This thesis relies on data-led analyses that leverage the breadth of the collected dataset to identify noteworthy phenomena in the practice of citing in international legal disputes. On this basis, the thesis offers four novel contributions to international law scholarship.

First, it investigates whet determines the “authoritativeness” of authorities in international law, i.e., what makes them authoritative as opposed to merely persuasive. Relevant factors include an authority’s age, author, or main topic. Second, it identifies the differences in the citation practices of different protagonists in international disputes, and attaches these differences to distinct strategic motives. Third, the thesis concludes that the central role of authorities international legal debates means that authorities circumscribe what is international law and what is not; in other words, the “citable” tends to delimit the boundaries of what is considered international law.

The fourth and last contribution lies in the Dataset collected to conduct these analyses, which includes nearly 200,000 citations to precedents and scholarly teachings from more than 7,000 documents in cases before the ICJ, the ITLOS, WTO panels, investment arbitration tribunals and the IUSCT. This unprecedented Dataset, and the methods used to gather it, will hopefully serve as a basis or as a blueprint for future empirical research in international dispute settlement.

This chapter explores how empirical methods drawn from data science - such as stylometry, network analysis, and linguistic forensics - can illuminate the involvement and influence of often overlooked actors in investor-state dispute settlement (ISDS). Traditionally perceived as a 'black box,' the inner workings of international tribunals, including the roles played by tribunal secretaries, institutional assistants, and other hidden contributors, remain opaque despite their significant impact on the legitimacy and authority of adjudicative outcomes. Leveraging computational tools such as Python-based text analysis, regular expressions, network analysis, and stylometric methods, this study empirically investigates the drafting roles of tribunal secretaries and assistants, challenging assumptions about authorship, delegation, and decision-making legitimacy in ISDS. The chapter particularly scrutinizes claims around ICSID's alleged streamlining of language in arbitration awards and presents empirical evidence illuminating these otherwise concealed practices. By introducing accessible empirical methodologies, this research underscores the value and necessity of data-oriented approaches to fully understand the dynamics of legitimacy and authorship in international arbitration.