Introduction to the research project ‘Legal Systems and Artificial Intelligence’

Responsibility: Principal Investigator (Japan) Mihoko Sumida

The aim of this project is to identify the social, economic and theoretical impacts and implications of the introduction of artificial intelligence (AI) into the Japanese and UK legal systems, particularly judicial decision-making. The project is funded by UK Research and Innovation (UKRI) and the Japanese Social Technology Promotion Organisation (JST) and is jointly conducted by the University of Cambridge (Centre for Business Research, Computer Laboratory and Faculty of Law) and Hitotsubashi University (Graduate School of Law and Graduate School of Business Administration).
  *“Human Information Technology Ecosystem(HITE)” is a research and development (R&D) focus area (Call for proposal Type) delivered by the Research Institute of Science and Technology for Society (RISTEX), Japan Science and Technology Agency (JST).

This research project is an ambitious challenge for the members of the Japanese team, namely Hitotsubashi University. Since its beginnings as a modern country, Japan has paid a great deal of attention to the development of law and legal systems, learning from Western countries, and has also boasted a high level of technological prowess in the field of industrial technology. However, LegalTech in Japan is still in its infancy, and various hurdles have been pointed out to the digitisation of the judiciary, including a delay in the digitisation of court documents. Together with our partners at the University of Cambridge, we are working on our research with the determination to overcome these hurdles and open up new horizons.

In other countries, including the UK, the use of machine learning (ML) to replicate aspects of legal decision making is already well advanced. A number of legal tech applications have been developed by law firms and commercial suppliers and are being used, among other things, to model litigation risk. Data analytics inform decisions on matters with legal consequences, such as probation, predictive policing and credit assessment. The next step would be to use machine learning to replicate the ‘judicial decisions’ core to the legal system, including adjudication. This project focuses here on the fundamental question: is law computable?

The duration of this research project is from January 2020 to December 2023.

Participants from Hitotsubashi University are Professor Mihoko Sumida, Professor Kazuhiko Yamamoto, Professor Keisuke Takeshita from the Graduate School of Law, Professor Yuichi Washida and Professor Mikiharu Noma from the Graduate School of Business Administration.
From Cambridge University, the participants are Professor Simon Deakin from the Centre for Business Research, Professor Jon Crowcroft from the Computer Laboratory, Senior Lecturer Dr. Felix Steffek from the Faculty of Law, Senior Research Associate Dr. Jat Singh from the Computer Laboratory and Trust and Technology Initiative, Research Associate and Affiliated Lecturer Dr. Jennifer Cobbe from the Computer Laboratory, and Research Associate Dr. Christopher Markou from the Faculty of Law.

Aims and objectives

There is an urgent need for an informed debate on the use of AI in the legal domain. This project aims to advance this debate by.

(i) Focusing on judicial decisions, which form the core of the legal system, the project will examine the possibilities, risks and limitations of their digitalisation and automation. As a research outcome, the project aims to develop elemental technologies and legal reasoning models that facilitate the introduction of AI into the legal system.
(ii) Explore stakeholders’ perceptions of the acceptability of AI-related technologies in the legal domain.
(iii) Develop guidelines to identify and address the legal and ethical risks associated with algorithmic judicial decision-making.


The project consists of three working packages, each introducing methods from horizon scanning (WP1), machine learning, deep learning, natural language processing and computational linguistics (WP2 and 3). In all these working packages, legal academics are also involved and work closely together to carry out research that is relevant to the operation of specific legal systems.

Working Package 1) Future Scenarios for Legal Systems and Artificial Intelligence Research Group: using the Horizon Scanning methodology (Team leader: Washida, Uehara, Sumida, Deakin, Markou)

The Horizon Scanning Method was developed principally by the Stanford Research Institute in the late 1960s. The method avoids the assumption that the future will tend to deviate from a linear extension of current circumstancesm, and attempts instead to develop more realistic predictions of the future by focusing on the collection and analysis of information that does not lie on the path of this linear extension. In implementing the Horizon Scanning approach we will firstly produce a database containing a range of information sources on the uses of AI in law, drawn from press reports and commentary and secondary academic literatures. The database will be used as the basis for discussion at a series of workshops. We will invite experts, researchers, corporate professionals and users across a broad range of fields of activity and different age ranges to take part in the workshops. Emergent scenarios will describe different possible combinations of advantages and risks stemming from the use of AI.

Working Package 2) Legal and Accounting Computation Research Group (Team leaders: Noma, Sumida, Takeshita, Deakin, Crowkroft, Singh, Cobbe, Markou)

This WP focuses on the normative requirements of legal norms, which are determined by comprehensively considering diverse factors, such as employee status, and examines whether the legal basis for these requirements can be visually represented using data from court cases and whether the outcome of a case can be accurately predicted using a decision-tree comprised of nodes corresponding to relevant legal indicators. This WP will use Deep Learning and NLP to analyse legal decisions for latent or hidden variables that can help inform and refine the model.

We will then explore how far the same techniques can be applied to the digitisation of knowledge systems used in accounting.

Working Package 3) Research Group on Prediction of Conflict Resolution by Artificial Intelligence (Team leaders: Yamamoto, Takeshita, Sumida, Steffek)

This WP deals with the prediction of dispute outcomes and aims to increase understanding of the problems arising from the use of artificial intelligence in predicting court outcomes by a range of stakeholders. The UK team is conducting research using a large dataset of UK court cases provided by the courts in England and Wales. Using this dataset, they will also test various machine learning approaches for predicting dispute outcomes. In order to conduct parallel research, the Japanese team has also received court case data from a Japanese court case database company, and is collaborating with the ‘Explainable Dispute Resolution Prediction Models for Civil Disputes’ research project (JST-ACT-X)* by Assistant Professor Hiroaki Yamada, Tokunaga Laboratory, School of Information Science and Technology, Tokyo Institute of Technology. Research is ongoing.

Furthermore, the WP plans to develop ‘Ethical Guidelines for Regulating Artificial Intelligence in Dispute Resolution’. The development of the guidelines will also be supported through round table meetings with representatives of the Japanese Ministry of Justice, the UK Ministry of Justice, the OECD Access to Justice Department, representatives of the Japanese and UK judicial communities and LawTech companies.

Progress and Results

Working Package 1) Future Scenarios of Law and Artificial Intelligence Research Group by Horizon Scanning

Both the Japanese and UK teams have generated social change hypotheses for 2030-2040 through the Horizon Scanning workshop, and have set the same theme of ‘the future of work and labour law’ as a future issue for further research. This will enable a comparison of future scenarios between Japan and the UK on how society will look like as the gig economy develops and AI is used for labour management and decision-making in companies. Ultimately, we plan to produce animation and other media based on the future scenarios that can be easily understood by the general public, and hold workshops with a variety of stakeholders to examine the differences between the UK and Japan.

Chikako Kanki, ‘The Potential Implementation of AI in Judicial Decision-making — From the Attempts to Determine Employee Status in the United Kingdom’. NBL No. 1187 (1 February 2021), pp. 12-20. [in Japanese]

Working Package 2) Legal and Accounting Computation Research Group

The UK team focuses on ‘worker’ status as a normative requirement and tries to visualise the basis for such judgments and elucidate their social and economic context. In addition to the determination of ‘worker’ status in contemporary labour law, the team is also building a dataset of historical employment cases, which will be used to test hypotheses about the long-term dynamics of legal change and the co-evolution of law with social and economic development. In the meantime, the UK team has also made progress in developing a conceptual framework, which has been published in the edited volume Is Law Computable? Critical Reflections on Law and Artificial Intelligence (November 2020, Hart/ published by Bloomsbury) and an academic article in the Journal of Cross-Disciplinary Research in Computational Law.

The Japanese team is focusing on the Forum Non Conveniens doctrine of international jurisdiction, developing an AI that implements the doctrine based on court cases from the UK, US and Japan, and conducting experiments to see whether the technology used makes a difference in accuracy. When this research project is successful, the UK team will be able to develop the second phase of research into a three-dimensional study by adding the perspective of comparison with the UK team’s work in the field of labour law.

Working Package 3) Research Group on Prediction of Dispute Resolution by Artificial Intelligence

The Japanese team is working on task design, large-scale annotation, building datasets and developing an explainable dispute resolution prediction system in collaboration with the research project “Explainable Dispute Resolution Prediction Models for Civil Disputes” by Assistant Professor Hiroaki Yamada in the Tokunaga Laboratory, School of Information Science and Technology, Tokyo Institute of Technology. As the first result of this work, the results of experiments conducted to ensure the accuracy of the annotations were summarised and presented.

The Japanese-UK team meets regularly online to share their findings with each other. In the second half of the project, the Japanese-UK team would like to try collaborative experiments.

In addition, the round table meeting with stakeholders, which was planned to be held by the UK team in Japan and the Japanese team in the UK, was changed to an online meeting, and students were invited to the online class to enable dialogue, thereby realising a novel online class that combines research with education and materials development. The online class was also changed to an online class, where students are invited to interact with each other. The intensive lecture “Technology and Legal Innovation” held at Hitotsubashi University in January 2021 was published in book form in March 2012. The book was published in March 2012.

Mihoko Sumida-Felix Steffek (eds.), Introduction to Legal Innovation, Kobundo, published in March 2022.[in Japanese] *cf)  news item in English,

The UK team built a large dataset of civil judgment documents provided by the courts in England and Wales and added machine learning-based analysis of 60,379 civil court cases on ‘contracts’ from 1709 to 2021.

◆ Research Cooperation:CourtCorrect(UK)