Enhancing Access to Justice via LLMs in the Field of Asset Division: CREA2 and the Digital Journey
Abstract
The ongoing digitalisation of both out-of-court and court services undoubtedlyenhances justice by increasing speed, lowering costs and time, providing informationand removing physical barriers to accessing legal services and enforcing rights. We contend that advances in large language models (LLMs) will further enhance access to justice by providing bespoke, data-driven services to citizens. In this paper, we concretise this claim by demonstrating how generative AI can be integrated into dispute resolution services through the example of our own project, CREA2, an online platform and digital tool for asset division between divorcing couples or beneficiaries of an inheritance. CREA2 was completed in June 2024, and in this paper we reflect on the finished project whilst looking ahead to the follow-up CREA3 project, which began in November 2024. This paper, moreover, seeks to address not only why LLMs are being implemented in dispute resolution, but also how. The implementation and evaluation of LLMs in legal practice is an inherently interdisciplinary endeavour. Therefore, here, drawing on both legal and technical expertise, we seek to situate the CREA project, and broader phenomena of AI-driven legal tools, within both the legal and technical discourses.
Downloads
Published
Issue
Section
License
EJLT is an open access journal, aiming to disseminate academic work and perspectives as widely as possible to the benefit of the author and the author’s readers. It is the assumption of the EJLT that authors who publish in the journal wish their work to be available as freely and as widely as possible through the open access publishing channel.
Authors who publish with EJLT will retain copyright and moral rights in the underlying work but will grant all users the rights to copy, store and print for non-commercial use copies of their work. Commercial mirroring may also be carried out with the consent of the journal. The work must remain as published – without redaction or editing – and must clearly state the identity of the author and the originating EJLT url of the article. Any commercial use of the author’s work - apart from mirroring - requires the permission of the author and any aspects of the article which are the property of EJLT (e.g. typographical format) requires permission from EJLT.
Authors can sometimes become no longer contactable (through, for example, death or retirement). If this occurs, any rights in the work will pass to the European Journal of Law and Technology which will continue to make the work available in as wide a manner as possible to achieve the aims of open access and ensuring that an author's work continues to be available. An author - or their estate - can recover these rights from EJLT by providing contact information.
The European Journal of Law and Technology holds rights in format, publication and dissemination.
EJLT, as a non-commercial organisation - which receives donations to allow it to continue publishing – must retain information on reader access to journal articles. This means that we will not give permission to mirror the journal unless we can be provided with full details as to reader access to each and every journal article. We prefer and encourage deep linking rather than mirroring. Encouragement is thus given for all users – commercial and non-commercial – to provide indexes and links to articles in the EJLT where the index or link points to the location of the article on the EJLT server, rather than to stored copies on other servers.
Please contact the European Journal of Law and Technology if you are in any doubt as to what this statement of use covers.