Contributions

Building a Commons of Free Access Legal Expertise: The Roles of LIIs

Abstract

The use of artificial intelligence (AI) in law, including in relation to decision-support systems, has again become a matter of great interest to both the legal profession (including NGOs involved in providing legal services) and to government organisations providing both legal advice and large-scale legal decision-making.  What might be sensible roles for free access legal information institutes (LIIs) in these new developments?

The previous wave of enthusiasm (and funding) for ‘AI and law’ from the early 1980s to the mid-1990s was to a large extent supplanted by the development of the World-Wide-Web and the provision of legal information via the web, of which the history of LIIs and the Free Access to Law Movement (FALM) formed a very significant part.

This paper first explores why the ‘Internet 2.0 context’ creates a very different environment from the pre-1995 situation of ‘AI and law’, and in particular why it makes it feasible to talk about the collaborative development of free legal advice services based on AI, in a way that was not previously feasible. 

The concept of ‘free access to law’ is not static, and the principles on which the concept is based require a response that includes the development of free access to legal advice, as part of ‘free access to law’.  ‘Freeing the law’ is a continuous process.

One of our conclusions is that LIIs are unlikely to be the builders of legal knowledge-bases in particular legal domains. They are much more likely to be the providers of tools by which such knowledge-bases are built, and the free access legal infrastructure within which they are built.

The final part of this paper set out the tools and infrastructure which the Australasian Legal Information Institute has developed (and is further modifying) to demonstrate ‘AI and law’ in the context of a LII.


ELI - Making Legislation Interoperable in Europe – From Dream to Reality

Abstract

In 2011, a group of governmental legislation publishers on the incentive of Luxembourg and in cooperation with the Publications Office of the EU, devised a system to structure and describe legislation data in a harmonised way. The objective was to make legislation more easily accessible and to allow interconnecting it: the European Legislation Identifier was born.

The concept resided on the premise that legislative interoperability can only be achieved if common standards are applied. The solution was designed to be flexible enough to maintain the legal traditions of different national legislation systems while providing a stable frame to enable interoperability

Concretely the solution combines three principles:

  • use of permanent URIs to guarantee stable, seamless and reliable access;
  • generation of descriptive metadata about the legislation to ease search;
  • dissemination of the metadata to enable structured reuse.
  • The success of the initiative depended on a number of factors, in particular the willingness of Member States to embrace it. This was achieved through the support of the Council of the European Union who created a dedicated Taskforce to drive the project forward. Council conclusions endorsing the ELI system, describing its nature and enshrining the principle of voluntary, optional and gradual implementation of ELI by national legislation publishers were published (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52017XG1222(02)). The ELI Taskforce – chaired by Luxembourg and is made up of all the ELI implementers, is responsible for the maintenance and evolution of the ELI standard. With the support of the Commission's ISA2 programme (https://ec.europa.eu/isa2/home_en) under coordination of the Publications Office of the EU, a set of tools and means to help the implementation of ELI was developed and can be accessed on the official ELI registry: http://eurlex.europa.eu/eli

    Today the dream has become reality and in 2018, 12 governmental legislation publishers, i.e. Austria, Belgium, Denmark, Finland, France, Ireland, Italy, Luxembourg (chair), Norway, Portugal, the United Kingdom as well as the EU Publications Office have adopted the ELI identifier for their legislation, while others are in the process of doing so. Non-governmental players from the legal or IT world are also increasingly showing their interest in ELI and the volume of elified legislation is steadily growing

    The ELI project will feature at this year's EU Datathon competition (https://publications.europa.eu/en/web/eudatathon/home), an event organised by the Publications Office of the EU which highlights the potential of linking EU and national data that will take place on 2 October 2018

    Today, the European Legislation Identifier is ready to be embraced as a reliable standard for cross-national legislation interoperability projects.

    More information about the status of the ELI initiative and how the dream has now become reality will be presented at the conference on 12 October.


Improving Public Access to Legislation Through Legal Citation Detection and Linking: The Linkoln Project at the Italian Senate

Abstract

Hyperlinks to cited norms, preferably resolved at the provision granularity level, are essential to improve readability of a legislative text. Besides for enhanced legislation navigation, the extraction and annotation of machine readable legal references as metadata of legislative texts guarantees interoperability and enables higher level applications in the Semantic Web and Linked Data domains. In Legal Information Retrieval, incoming and outgoing legal references of documents can be exploited to improve search results.

In order to overcome the limitations of the existing tools for the automatic extraction of legal references from Italian legal texts, in 2015 the Italian Senate promoted the design of a new software that could cover a wide number of authorities and typologies of act, support the main standard identifiers for legislative texts, be efficient, maintainable and easy to integrate in different environments, like web applications, existing legal platforms and so on. Hence, ITTIG developed Linkoln, a Java project, that was later integrated in ShowDoc, the application for the visualization of official acts (including legislative proposals, amendments, dossiers, etc.) on the public website of the Italian Senate.

In order to tackle the complexity of automatic legal reference extraction, Linkoln relies on a pipeline of specific services that analyze the text in order to identify, normalize and annotate the relevant textual features of a reference. Implemented using JFlex, a popular lexical scanner generator for Java, these annotation services support macros of regular expressions, rules and state, compiled into efficient lexical automata. At the end of the pipeline, patterns of identified entities are annotated as reference objects anchored to the original text. Finally, Linkoln associates one or more standard identifiers for legislative texts and URLs to the found references. Currently urn:nir, celex and eli are supported.

Besides plain texts, Linkoln also accepts previously annotated texts as input (HTML, XML) and it is able to render the additional reference annotations while preserving any pre-existing annotation. ShowDoc exploits this capability allowing users to detect legislative citations and activate hyperlinks while reading a document or a document partition by invoking Linkoln on the currently visualized HTML fragment. In this scenario Linkoln receives previously marked-up HTML text as input and returns the enriched HTML annotation with <a> tags in correspondence with the detected legislative references.

With the Linkoln project, we made available in the public domain a robust and reliable framework for the automatic detection of legislative references from Italian legal texts. After the adoption and integration in the publication workflow of the Italian Senate and thanks to its compliance with (both legal and web) open standards and its release as open source software, several Italian enacting authorities and Public Administrations, will be encouraged to test and adopt the software and contribute to its iterative refinement, evolution and maintenance.


Law As Data: The Implications for Government Law Drafting and Publishing Offices

Abstract

The collection, dissemination and elaboration of law by non-government actors needs to be influenced by, and in turn influences, the making and publication of law by government.

The LVI2018 Call for Papers notes that “the rise of computational power and development of  the internet … is impacting the way in which legal information is collected, disseminated and elaborated.”

That statement rightly recognises the changes taking place in the way that, once law is made and published by Government, for-profit and not-for-profit organisations process it (i.e. collect, disseminate, elaborate).

This paper is written from the perspective of a government law drafting and publishing office and asks the subsequent question: “what implications do those changes have for the way government bodies should make law and make it available?”.  These expectations on government are driven not just by the fact that law is data, but also by the constitutional expectation that in a state based on the rule of law citizens should have ready access to the law.

Changes in the way government makes law and makes it available will in turn change the underlying data (i.e. law) which for-profit and not-for-profit organisations will process.

Law making and publishing by government on the one hand, and processing by non-government bodies on the other, are therefore part of an interconnected system. The changes in both parts are likely to be iterative, interdependent, diverse and unpredictable.

It is therefore necessary for all involved in the system to have an understanding of the changes and to keep abreast of developments across both the governmental and non-governmental parts of the system.

From the perspective of the New Zealand Government’s lead law drafting and publishing office this paper reviews current relevant developments in New Zealand, and asks the following questions, illustrated by current examples:

In what circumstances should law be made available (or even made) as code?

What implications do these developments have for the content of law?

What are the implications for Government law publishers of the requirement to be good data custodians?


Publication of Law for the Digital Age

Abstract

The present promulgation of laws and judgements must change in the digital age. Therefore, “authors” and “publishers” of legal information have to consider a more comprehensive approach. We propose a quite revolutionary setting: publication in text should be supplemented by visual as well as logic-ontological representation. As laws are more and more implemented by software agents and robots, parliaments have to consider publishing laws also in a logic-ontological representation.

We propose that law should be presented in three levels: text, visualization and logic-ontological code. These three tiers constitute an inter-connected communication of law. Whereas the representations are functionally independent, they should communicate the same content. The representations are strongly interrelated by links between same rules or concepts. Additional representations, e.g. another language, or another level of the same language, can be easily added to this representation method. The only legally binding form should be the textual representation. However, the other forms of representation should have the presumption of legal certainty. The publication of the other levels depends on developing a mostly automated system of generating a logical and ontological representation; re-using some support by parliaments and courts (e.g. semantic mark-up). The logical representation refers to the work of Hajime Yoshino on Compound Predicate Formula and our own work on legal knowledge representation.

The results are preliminary but the understanding of laws is much improved if other representations are also available. Laypersons can use software agents and apps to learn about applicable legal rules.


The EU Council Conclusions on the On-line Publication of Court Decisions

Abstract

On 8 March 2018 the Council of the European Union decided upon the ‘Conclusions of the Council and the representatives of the Governments of the Member States meeting within the Council on Best Practices regarding the On-line Publication of Court Decisions’ (hereinafter referred to as ‘the OPCD Conclusions’). This contribution discusses history, status, context and contents of this document.

The OPCD Conclusions contain 17 recitals, grouped in five sections. The introductory section describes the general incentives for publication of court decisions and stresses the special need within the European Union for knowledge about the legal systems of all Member States. In this section it is also explained that these conclusions should be viewed upon as best practices that do not strive for harmonisation but merely are an invitation for reflection and inspiration.

The second section is on selection. Given the wide variety of practices within the Member States regarding whether all decisions have to be published or just a selection, the OPCD Conclusions do not state any preference. However, they do express that if a selection is made criteria should be drafted and published, to give guidance to the publishing courts and offer transparency to the public.

The next section is about data protection, also a topic on which policies between Member States diverge. Recital 10 permits not to obscure personal data at all, but ‘Member States are called upon to consider the implications of […] the General Data Protection Regulation.’

The fourth section is on reuse. Member States are invited to make published decisions and their metadata – given technical and budgetary constraints and the features of the drafting process –available for re-use in computer-readable formats. 

The final section is on improving usability. One of the issues highlighted is the problem that voluminous repositories create the needle-in-the-haystack problem. To facilitate separating the wheat from the chaff it is advised: “To supply for some kind of importance qualification, indicating which, and to which extent decisions are of relevance for others than the parties to the case.” This provision can be considered to be the most innovative of the OPCD Conclusions; it requires explicit action – by manual tagging or developing automated means – to improve the accessibility of published decisions.

‘Council conclusions’ are a policy instrument by which the Council of the EU can express a political position or commitment when it is not exercising the powers conferred upon it by the Treaties. The Council decides on these conclusions by consensus only. Council conclusions do not have legal effect, but often can be classified as soft law. Since the OPCD Conclusions are the first official EU document on the on-line publication of national court decisions, they can be expected to have an impact on how courts and court administrations will judge their own performance and shape their ambitions.


The European Parliament's Digital Transformation Journey and Its Impact on Law Accessibility and Participation

Abstract

The European Parliament has a consolidated experience in promoting the widest access possible to the legal texts adopted at the EU level. The increased services offered to citizen’s in accessing not only the adopted texts, but the also complete decision-making process are a testimony of the intent to strengthen the bonds between the Institution and the citizen through the use of ICT.

The Directorate-General for Innovation and Technological Support (DG ITEC) is a central actor in promoting and implementing the digital transformation of the Parliament. The level of maturity of the digital journey undertaken by the European Parliament enables to conceptualize the digital relationship with the citizen around three objectives that will be the central topic of the presentation.

The first objective is to digitalise the legislative chain in order to create one single, seamless process that goes from the decision-making to the publication. The e-Parliament Program is the series of IT projects that support this ambitious objective. Throughout the last years, e-Parliament has delivered many solutions (on amendments, legal texts, verbatims, etc.) that will be outlined in the presentation.

The second objective responds to the need of reducing the time elapsing between the production of the document and its availability to the public.  An Institution that wants to be closer to the citizens has to factor in this variable of transparency, and DG ITEC has made substantial efforts in that direction through a wide xmlisation effort.

The third objective is participation. Transparency has conceptually shifted from “access” to “participation” in the recent years, and this impacts the ways a Parliament work. The European Parliament launched several initiatives in cooperation with other Institutions and also the EU Publication’s Office, the most notable being the Linked-Open Data project providing the technical solution to a multi-lingual, in real time, participative process on the EU laws. Two other examples supporting this shift are the ePetitions Portal and the Online Consultations initiatives carried out at Parliamentary Committee level. Both initiatives will be illustrated during the presentation to illustrate the European Parliament’s approach to participation from a technological point of view. 

Finally, the presentation will put all these innovative solutions to diversify the ways of interacting with the citizens in the perspective of the broader strategy of reinforcing democracy through technology.  The reflections on this topic have been divulgated through several publications these last years that will be referred to during the discussion.


The Role and the Foreseeable Future of the Library of the Court of Justice of the European Union

Abstract

The free online catalogue of the Library of the Court of justice of the European Union gives access to all its bibliographical records and allows everyone who has an internet access to make effective researches on EU law and other fields of law.

We will analyze the tasks of the Library and the collections in order to find out how this could facilitate access to legal information in certain fields (A) and then we will examine the changes that the library is facing, due to the technological development, to the use of the electronic resources and to the more and more stringent constraints on financial and human resources (B). Within the mission there are some additional functions of the Library that are complementary to the main goal – which is to be a depository for all the documents on European integration and the promotion of knowledge of the European legal system, performed by providing an online catalogue, a current bibliography, information on recent acquisitions and, once a year, a juridical bibliography on European integration and by giving access to the library to everyone who can manifest a strong interest in EU-law and related areas.

Auxiliary objectives

The strengthening of the cooperation with other libraries (the Eurolib network, National Library of Luxembourg, national judicial libraries, academic and university libraries, particularly in the area of law) should allow an improvement into the Library’s efficiency, especially with the exchange of data and of documents. It should be also a way to increase the awareness of the services offered by the Library and its online catalogue.

A reorganization of the service could also further the developing of new opportunities and initiatives related to vocational training and improve the communication between members of Library staff.

 

Conclusions

To bring into focus the mission of the Library, it is necessary to collect all documents reflecting the academic studies on the European integration process both in paper and digital versions. Despite limitations imposed by copyrights and related rights, and tendencies to tighten the conditions of licenses by publishers, the Library of the Court of Justice is expected to keep in mind a universal rule underlying the activity of every library, namely widespread and free access to knowledge.


A Future between Open Access and Artificial Intelligence: Which Economic Model for Legal Publishing in Belgium?

Abstract

Federal Belgium offers multiple legal orders and as many official sources of law. This multilingual and multifaceted legal environment makes the analysis of the legal publishing market and its par excellence « captive audiences » interesting, following one possible scenario, especially for the French-speaking part of Belgium. We no longer speak of the « computerization of law » but rather of an additional stage in this computerization: the introduction of artificial intelligence whose application to law is based mainly on the use of two techniques, natural language processing and machine learning. What does this have to do with publishing? The link is very simple as “Doctrine” equates to unstructured information.

Under the impact of a weakening of their economic model confronted with information and communication technologies as well as the Open Access movement, legal « publishers » are moving towards services with high added value as they still control in an important way the three main sources of law (legislation, case law and doctrine). Publishers want to continue to play a decisive and deciding role in the « legal market ».

For several years now, major legal publishers have been investing in artificial intelligence solutions  and rely in part on the acquired capital and the remanence of the conventional editorial process.

The strong editorial control on the legislative and regulatory texts consolidation process, the possession of exclusive collections of case law as well as the hold over doctrine – here understood as a « captive public » – given the control of publishers over a far from negligible part of the editorial process and management of doctrinal publications, make available a material on which new applications of information technology can be built. Thus, it would seem that the publishing, at least part of it, of a certain doctrine only becomes the essential supplement of information to « refine ontologies » set after having recourse to data mining within databases in which the information is structured.

The challenge is to structure the « legal information » represented by the « edited doctrine », since we are in the presence of sources containing information that is not structured and that is recorded in natural language. It is easy to imagine – and the doctrine has already asked itself the question – that the language and writing of the law, not only that of legislation and jurisprudence but that of doctrine, are in a hurry to adapt in order to ensure that the data are « cleaned and adjusted » so that the machine can obtain better analytical results thanks to the technique that is text mining.

Our contribution will aim to show the slow but certain evolution of the economic model within a well-defined publishing framework but influenced by wider phenomena.


Information Exchange in the Big Data Era

Abstract

The enormously increase of data availability poses new challenges and suggests new interesting road to  public and private data produces and providers.

The European Commission acknowledged these opportunities which can significantly boost European competitiveness in the global market and in scientific research. Very recently, the EU has created a solid framework for digital trust, through the publication of the new General Data Protection Regulation, which regulates the free movement of personal data, based on trusted and widely accepted data technologies.

One of the cornerstones of the building process of a common European data space

is the possibility to access and share public and publicly funded data.

This complex tasks has many important different goals. Here we mention a few of them:

  1. Citizens’ secure access to and sharing of health data
  2. Improving and innovating healthcare solutions, based on tele-medicine and mobile applications, in full compliance with the current data protection legislation.
  3. Multiple and different uses of public sector information. In fact, public sector and National Agencies produce a huge quantity of information in terms of raw data which can be combined in order to produce innovative digital services and organize better policy making. To this end, the European Commission has also proposed the creation of a new Support Centre for Data Sharing, to be established in 2019.
  4. Sharing Scientific Information, in order to facilitate the dissemination of results and to enhance the creation od scientific networks acrosso countries.
  5. Economics: Business to Business (B2B) data sharing, which should consider the availability of “non personal machine-generated data” under the principles of transparency, shared value creation and the avoidance of distorted competition.

Even more interesting and innovative, however, the General Data Protection Regulation enhances and supports the access to private sector data for public interest purposes (B2G).  Data held by private companies, such as supermarket chains, telecommunications operators, car manufacturers, social media or online platforms could be of enormous guidance in better urban planning, market monitoring, traffic management and many other fields.

All these new challenges suggest a list of exceptional problems to be faced both on a legislative and on a methodological grounds


The Bibliography of the Italian Parliament: Enhancing and Disseminating a Bibliography Using the Linked Open Data Model

Abstract

The Bibliography of the Italian Parliament and Electoral Studies, also known as BPR, is a database created and mantained by the Library of the Chamber of Deputies and freely available on the Internet at http://bpr.camera.it and from within the historical portal of the Chamber of Deputies (http://storia.camera.it/bpr#nav).

It currently contains nearly 22,500 bibliographic references relating to the Italian Parliament, the post-1945 National Consultative Assembly and Constituent Assembly, and the general elections, since 1848.

Each reference is assigned one or more classification codes, which are taken from a directory of over 100 classifications that are organized into seven major fields: General topics (A), Institutional relations (B), Sources of law and organizational aspects (C), The lawmaking function (D), Parliamentary oversight, policy-setting and information (E), Members of Parliament (legal and social status) (F), Elections (H).

The BPR is also a digital library: it offers more than 3,000 full texts publicly available, and other 1,000 are accessible only through the Intranet of the Chamber of Deputies and the Senate.

In the 2010s, the BPR underwent a major restructuring process whose main objectives were: to substitute the original Access database and to add more management features; to redesign and update the end user interface, making it more agile and including faceted search; to increase the external projection of the BPR, permitting data harvesting and general data reuse. The project was carried out jointly by the Library, the Information Technology Department and a contracted company and achieved its first results in 2016.

As part of the more general adoption of linked open data at the Chamber of Deputies, it was decided to adopt the linked open data model both internally, using a modular system in native RDF format to manage citations, and in the publication of data, not only on the official data portal of the Chamber of Deputies, but also on the new BPR website.

The whole process can be described as an iterative incremental life cycle model, structured in five main activities: specification (analysis of the data sources, URI design, definition of the license); modelling of the ontology; generation (transformation, data cleansing, linking); publication; exploitation.

The focus is currently on adding links and enhancing discovery. Several datasets were considered to be suitable as linking targets: firstly, the parliamentary biographical datasets already published by the Chamber of Deputies, secondly, external datasets such as VIAF for name authorities, and the Italian Nuovo soggettario for classification but also more general information resources such as Wikipedia and its linked data equivalent DBpedia.

A very important source for the compilation of the BPR has always been the DOGI database, maintained by the CNR Institute of Legal Information Theory and Techniques. Born in 1969, it provides abstracts of articles published in more than 250 Italian legal periodicals. Since 2017, the BPR working group has been receiving a bimonthly selection of updates from the DOGI database, so that it can concentrate on going through other periodicals and miscellaneous works. Some hypothesis are made on further developments of this cooperation.


The Right to Science and Open Access to Legal Knowledge: An International Law(yer) Perspective

Abstract

The paper aims to addressing the issue of the dissemination of the results of scientific research, and in particular of legal scholarship, in an international/EU law perspective.

The international legal order contains the definition of the right to science in the art. 27 of the Universal Declaration of Human Rights; the same right, repeatedly cited in several acts adopted by the United Nations, is also included in binding universal and regional instruments and, in the form of the principle of sharing the benefits deriving from scientific research, is also applied in international treaties pertaining specific issues.

Once affirmed the existence of a right to science in contemporary international law, less unambiguous is the reconstruction of both its nature and content: while some authors conceive it as an independent right, that adds to other fundamental rights and deserves an autonomous protection inasmuch, as it is aimed at increasing the material and spiritual quality of the life of individuals and collectivities, other scholars build it as a mere instrument for the implementation of “classic” fundamental rights such as culture or health.

Among the main concrete applications of the right to science there’s the one related to the free dissemination of research results, promoted in the scientific community by the well known “open access” movement, whose principles were defined in two declarations, adopted respectively in Budapest in 2002 and in Berlin in 2003. It is even more relevant in for the quest of the open access for the best ways to disseminate the results of public funded research: one may think, for instance, to how the dissemination of these results through publications distributed by private companies asking fees could infringe the right of access to science itself.

Despite the aforementioned widespread diffusion in the international legal order, the concrete scope of the right to access to scientific knowledge, however, only recently was addressed by the international legal scholarship: it’s time that international legal doctrine, looks into the mirror and starts the quest for the right of its own diffusion.


Towards Open Science and Open Doctrine. The Principles and Laws That Promote the Sharing of Knowledge, and How They Are Implemented

Abstract

The evolution of open science in France is almost completely the result of constant friction with the business models that drive major international publishing houses, where each party has adapted to developments introduced by the other, but also of practical steps taken to ensure that shared documents are efficiently collected and made accessible. We will provide several examples of the development of Open Science in France, such as the platform http://dissem.in.

How open science principles are effectively implemented in the area of legal knowledge in France. What can be done to encourage law scholars to publish their work on a single common platform? And which platform should that be? Should it be improved, and, if so, in what way? Will dialogue resolve conflicts and pave the way for open science in a viable economic context?


A Model of Justice As a Platform: A Case Study of Open Data Disclosure

Abstract

From several years, digital technologies are influencing operations and managerial processes in Courts of Justice.

Obviously, digital technologies’ impact is highly diversified across Europe following national procedural norms and the degree of autonomy of Courts of Justice within institutional arrangements.

Notwithstanding peculiarities, Courts are steadily incrementing the amount of available data which represent a huge resource that deserves to be capitalized.

The research wants to deepen the topic of how to make data available to Courts organizational environment.

The theoretical and methodological framework stems from crucial assumptions.

First, datification and information gathering are core processes shaping contemporaneity. This is true not only in the for-profit markets where platform-shaped companies have acquired a leading role, but also in the definition of public policies implementation networks that are increasingly based on sharing and managing data.

Second, provision of judiciary services produces a platform providing a deep understanding of social and economic systems. Decisions issued by a court have a profound impact on highly diversified social areas, including for example assistance of people who are non-autonomous or excluded from social life.

Enabling justice as a platform in which different subjects can acquire crucial information and data becomes a leading objective in particular for the elaboration of public policies, but also to empower social innovation and entrepreneurial opportunities.

Third, the care of the relationship with stakeholders and their active engagement is essential for Courts effectiveness. Far from being just a communicative strategy, relationship with stakeholders can become an important driver for efficiency improvement of internal processes.

The research will offer an empirical case study describing a project of open data disclosure carried out in Torino Court of Justice civil sector with collaborative methodologies.

The case study enables a deeper understanding of what it means to move towards a public disclosure of data. It has been a privileged opportunity to observe closely cultural and cognitive framing of open data by administrative clerks, the main technical obstacles posed by Italian Courts’ digital system, and exchanges of data already taking place with other administrations.

The main goal of the research is highlighting key areas of disclosure and their impact on Courts of Justice environment, the main drivers of innovation and the most important limits to be overcome in order to fully enable justice as a data and social platform.


CREA - Conflict Resolution with Equitative Algorithms – A Cloud-based Decision Support System

Abstract

Efficiency of justice is one main pillar of the right to a fair trial, held in Article 6 of the European Convention on Human Rights. Reasonable time is one of the most frequently invoked components of this efficiency, but, at present, the gold standard of having a day in court for getting a judgment in a cost effective and timely way is not feasible anymore. The proper functioning of cross-border

and national civil proceedings in European Member States is on the edge.

This work aims to introduce new forms of access to justice, opting for a privatised method of dispute resolution, such as helping tools for lawyers, negotiators, mediators and judges, to guide parties into reaching a fair agreement, by applying algorithmic procedures - largely already studied from a mathematical and economical point of view - and artificial intelligence methods into the legal field, safeguarding minimum procedural guarantees.

As a result, we propose a distributed and efficient cloud-based decision support system, CREA - Conflict Resolution with Equitative Algorithms - by implementing all decision support and analysis functionalities as services for the parties, acting as main characters of the final settlement. CREA has been awarded from the European Commission with a financing for a two years’ work. The project has started in October 2017.


Digitalised Legal Information: Towards a New Publication Model

Abstract

This presentation will outline some of the key developments regarding publication and communication of legal rules and standards (i.e. legal information) to show that efficiency and sustainability of dissemination of legal information is reliant on how we design the entire model of publication of legal information.

My original claim is that future work shall not focus only on the (new) methods of publication and communication of legal information. Methods of publication represent merely a single aspect of a holistic publication model, and they must be seen as such. I will argue that in light of the digital challenge, i.e. in the light of digitalisation of legal information, big data analytics, AI, etc., it is appropriate to revise the entire publication model and not only some aspects of it. An addressee-centric publication model will be outlined as a prospective solution to problems posed by the digital challenge. In this model, a relevant subset of digitalised legal information is delivered to its intended addressee in a similar way as targeted online advertising. Unlike the existing research that promotes personalisation of law (personalised legal information), my presentation will advocate personalisation of publication models.

To support my argument, I will analyse models of publication as they appeared in the prehistorical, historical, as well as hyperhistorical stage of human evolution. This will help me to demonstrate how legal information was delivered to its intended addressees, i.e. to those who were expected to obey the published laws. I will show that the emergence and developments of ICTs shaped significantly the models of publication of legal information. In doing so, I will pick some paradigmatic models of publication from each of the three stages of societal evolution. to show that the progress regarding publication models was driven by efficiency and sustainability considerations.

Said overview of these models will also help me to highlight four main elements that all of these publication prototypes had. In particular, I will point out that each model must have a subject-matter of publication, a method of publication, a publisher of legal information, and an addressee of published or communicated legal information. Distinguishing the four elements will support the view that ICTs represent only a method of publication, i.e. only a partial aspect of a more complex problem. An entire publication model, in contrast, represents a holistic description of a system in relation to which we can better assess efficiency and sustainability of dissemination and communication of legal information. This insight is crucial for our assessments and revisions of the existing publication models as well as of various methods of publication.

Finally, I will canvass the currently prevailing EU model of publication and I will consider what are the implications of the digital challenge in relation to all four aspects of the current model (subject-matter; method; publisher; addressee). This will allow me to outline minimal features of a new, presumably more efficient, and better sustainable model of publication of digitalised law.


Dissemination of Legal Information: Wedding or Divorce between Open Data Movement and Implementation of Personal Data Protection Law Principles

Abstract

The goal of this proposal is to analyze if open data law principles and personal data protection law principles can be applied simultaneously in case-law publication and if yes, how can we implement both of these principles : what type of data have to be made available to the public and what type of data have to be processed, according to the personal data principles?

The first question to address is to know what can be considered as open/public data. Based on the personal data law principles, name of parties have to be processed but how. Another question comes up with the surname is the gender because some names reveal it sometimes. Should we add Mr. or Mrs. before to respect the gender revealed by the name and surname or does it reveal too much about another personal data: sex? What about additional details such as parties’ post address, date and place of birth?

As it prescribes by recital 14 of the General Data Protection Regulation (G.D.P.R.), personal data regarding legal persons are not covered by this regulation so they do not require to be processed carefully and be protected against unauthorized uses. However, it is more questionable when company name or legal status reflects the Company Executive Officer (C.E.O.) identity. Should we interpret that type of data are strictly related to legal persons or should we apply the same rules for natural persons’ personal data and replace part of company name containing name and/or surname of the C.E.O. by their initials as well ? I will share my experience as an author for a database about Intellectual Property case-laws in South-East Region in France, so-called «R.P.I.S.E.» from January 2015 to June 2017 to answer these questions.

According to the French access to administrative documents legislation, public data which contain personal data can be reused in three cases. Based on the personal data principles, there are differences between anonymization and pseudonymization. The G29 Working Group indicate three criteria to ensure that data are anonymized.


Open Legislative Data and Its Impact on Lobbying

Abstract

The contemporary legislative process, in the context of the information society and the rule of law, creates a new task to parliaments and regulatory agencies: the need to communicate efficiently not only with human beings, but also with machines. Today, transparency is not limited to the mere provision of information on the internet. Is necessary to generate data in open formats, consumable by computers in order to be easily reused in digital applications.

Corporations have been using algorithms and big data to solve some of their biggest challenges. Those tools are now starting to catch on in lobbying. Today, several technologies collects data from many sources and compiles it in a searchable database where is possible to add an analytical layer.

The aim of the session is to show how the legislative openness are impacting the decision-making process and what are the challenges that Parliaments will need to face to reduce inequalities related to the participation of different social actors.

To build a better government, we may just need better lobbyists. This sentence may be considered extremely simplistic, at first, but it gains prominence when lobbying is viewed as a practice of political participation on building more efficient and effective State decisions. On the other hand, talking about “better lobbyists” may be the same as talking about high-quality civic participation, broad, conscious and transparent, unrelated to the vicissitudes and crimes that sometimes ends up contaminating the legislative process.

Few high-cost firms offers technologies for lobbying purposes. Besides this, Data Science is an area dominated by a limited number of social actors, notably the most economically privileged. Therefore, the mere disclosure of open legislative data can contribute to the accentuation of social inequality. So, Parliaments and also the Executive Branch needs to rethink its openness values, acting as an inducer of the co-creation of technologies that could help to solve inequalities on the legislative decision-making processes.


Datificazione delle relazioni socio-economiche e diritto comparato / Datification of Socio-economic Relationships and Comparative Law

Abstract

Grazie agli smartphone, ai social media e alle reti di sensori, oggi nel mondo sono generati più dati che mai. La tecnologia può aiutare a dare un senso a questi dati per migliorare le nostre vite. Questa è la data-driven innovation.

La raccolta, l'archiviazione e l'analisi dei dati stimolano l'innovazione in molti settori della nostra società e dell'economia, tra cui sanità, istruzione, trasporti, comunicazione, economia e sicurezza.

Per le imprese, l'analisi e l'elaborazione dei dati offrono nuove opportunità, trasformando in meglio efficienza e produttività. Per il settore pubblico, un migliore utilizzo dei dati consente servizi più efficienti, trasparenti e personalizzati. Per gli scienziati, open data e open results consentono nuovi modi di condividere, confrontare e scoprire nuovi campi di ricerca. Per i cittadini, i dati sono portatori di maggiori informazioni, per servizi e applicazioni nuovi e migliori.

In questo quadro, dal punto di vista giuridico, si pongono questioni in tema di titolarità, circolazione e accesso ai dati, ed i legislatori sono in cerca di strumenti efficaci per garantire diritti fondamentali come la privacy, la sicurezza, il benessere del consumatore e la concorrenza.

Il fenomeno della datificazione delle relazioni socio-economiche si interseca pertanto con il diritto comparato, sotto vari profili. In primo luogo, la prospettiva comparatistica conduce a individuare le caratteristiche peculiari dei dati attraverso un inquadramento concettuale dei medesimi nell’ottica di altre discipline, in particolare l’economia e l’informatica. Inoltre, in considerazione delle questioni giuridiche suddette, il diritto comparato risulta essenziale al fine di garantire che i servizi di gestione e analisi dei dati possano essere sviluppati e forniti oltre le frontiere nazionali. Infine, d’altro canto, l'applicazione dei metodi di analisi dei dati alle questioni giuridiche potrebbe generare un consenso più ampio e le leggi di diversi Paesi potrebbero essere considerate come dati da analizzare ed elaborare per osservare schemi e percorsi evolutivi del diritto.

Thanks to smartphones, social media, and sensor networks, more data about our world is generated today than ever. Technology can help make sense of these data to improve our lives. This is data-driven innovation.

Data collection, storage, and analysis drives innovation in many fields of our society and economy, including healthcare, education, transportation, communication, economics and security.

For companies, data analytics and processing bring new opportunities, transforming efficiency and productivity. For the public sector, better use of data allows services that are more efficient, transparent and personalised. For scientists, open data and open results allow new ways to share, compare, and discover whole new fields of research. For citizens, data brings more information and empowerment, and to new and better services and applications.

In this framework, legal challenges arise about data ownership, circulation and access to data, and legislators are trying to find effective ways to guarantee fundamental rights such as privacy, security, consumer welfare and competition.

The phenomenon of the datification of socio-economic relationships therefore intersects with comparative law, under various profiles. First of all, the comparative approach leads to identify the peculiar characteristics of the data through a conceptual framework of the same in the perspective of other disciplines, in particular economics and information technology. Moreover, considering the above-mentioned legal issues, comparative law is essential to ensure that data management and analysis services can be developed and provided across national borders. Finally, on the other hand, the application of data analysis methods to the legal questions could generate broader consensus and laws from different countries could be considered as data that needs to be analysed and processed to observe patterns and evolutionary paths of law.


Fra data governance e protezione dei dati: la dimensione giuridica dei big data nella sfera pubblica / Between Data Governance and Data Protection: The Legal Dimension of Big Data in the Public Sphere

Abstract

Nel contributo gli autori intendono proporre alcune riflessioni sulla dimensione giuridica dei big data, con particolare riferimento all’ambito pubblico.

Nella natura stessa dei big data emerge un profondo mutamento nel modo di conoscere e misurare la realtà, basato adesso non sul nesso causa/effetto, ma su inferenze e correlazioni spesso opache. Nella stessa analisi algoritmica e nel correlato processo di estrazione della conoscenza si annida, così, anche il rischio di giungere a conclusioni imprecise o discriminatorie, frutto di errori tecnici o manipolazioni deliberate.

I big data sollevano questioni estremamente complesse sul piano giuridico.

La loro stessa “proprietà”, ossia l’attribuzione della titolarità sui big data, è problematica, dal momento che esistono legittimi titolari dei dati e dei dataset che compongono i big data, ma le stesse raccolte di big data sono ascrivibili a banche dati con la relativa tutela del diritto d’autore o del diritto sui generis, a seconda delle caratteristiche e della correlata afferenza.

A tale profilo si collega strettamente l’individuazione dei profili di responsabilità, soprattutto quando nella gestione degli stessi intervengano più soggetti.

L’utilizzo dei big data e la knowledge ad essi correlata generano importanti implicazioni etico-sociali che devono essere oggetto di particolare attenzione in ambito pubblico, in quanto foriere di possibili asimmetrie informative, di squilibri di potere e di un vero e proprio big data divide con la collettività di riferimento. Il rischio si sostanzia nella possibilità di arrivare a un controllo sociale e a una sorveglianza totali grazie a un uso dei big data per finalità estranee a quelle dichiarate e comunque ignorate dagli utenti.

Proprio la limitazione delle finalità di raccolta dei dati e la necessaria trasparenza delle informazioni nei confronti soggetti interessati sono due aspetti della protezione dei dati personali più esposti ai rischi di frizione fra la gestione dei big data e i principi della data protection.

L’utilizzo dei big data metta alla prova alcuni cardini del regolamento europeo 2016/679, quali i principi di limitazione della finalità, di minimizzazione e di esattezza dei dati, la disciplina delle informazioni sul trattamento e del consenso. I problemi sono acuiti dalla difficoltà di garantire una reale anonimizzazione, a causa dei rischi di re-identificazione del soggetto e, altresì, della definizione stessa di dato personale, che nel contesto dei big data può risultare non adeguata rispetto alla protezione del singolo.

Accanto agli elementi che ancora sembrano stridere con una tutela efficace, il regolamento europeo contiene elementi che, invece, possono essere opportunamente valorizzati nel contesto dei big data, che mirano a una “privacy tecnica” e a un correlato approccio preventivo del rischio sulla data protection, basato sull’accountability e sull’efficacia del momento sanzionatorio.

Accanto alle problematiche giuridiche, l’utilizzo dei big data in ambito pubblico impone di porre attenzione al profilo della governance degli stessi: di dati, infatti, si occupano diverse autorità (l’Agenzia per l’Italia Digitale, il Garante per la protezione dei dati personali e l’ANAC).

Nelle conclusioni si tenterà di evidenziare le potenzialità di alcuni strumenti di gestione dei big data e, di conseguenza, saranno proposte alcune strategie per una gestione dei big data in ambito pubblico che sia idonea a tutelare i diritti umani e le libertà fondamentali della persona.

In the contribution the authors intend to propose some reflections on the legal dimension of big data, with particular reference to the public sphere.

In the nature of big data emerges a profound change in the way of knowing and measuring reality, based now not on the cause / effect nexus, but on often opaque inferences and correlations. In the same algorithmic analysis and in the correlated process of knowledge extraction there is also the risk of arriving at inaccurate or discriminatory conclusions, as a result of technical errors or deliberate manipulation.

Big data raise extremely complex legal issues.

Their “ownership”, i.e. the attribution of ownership over big data, is problematic, since there are legitimate data owners, as well as owners of the datasets that make up big data, but the same collections of big data are attributable to databases with the relative protection of copyright or sui generis right, depending on the characteristics .

This aspect is closely linked to the identification of the profiles of responsibility, especially when more subjects are involved in their management.

The use of big data and the related knowledge generate, in fact, important ethical and social implications that must be the subject of particular attention in the public sphere, as they make possible information asymmetries, power imbalances and a real big data divide with the reference community. The risk consists in the possibility of achieving total social control and surveillance thanks to the use of big data for purposes unrelated to those declared and however ignored by users.

The limitation of the purpose of data collection and the necessary transparency of information to stakeholders are two aspects of the protection of personal data most exposed to the risks of conflict between the management of big data and the principles of data protection.

The use of big data tests some of the cornerstones of the European regulation 2016/679, such as the principles of limitation of purpose, minimization and accuracy of data , the regulation of information on processing and consent. The problems increase because of the difficulty of guaranteeing a real anonymity, due to the risks of re-identification of the subject and, also, of the definition of personal data, which in the context of big data may not be appropriate with respect to the protection of the individual.

Alongside the elements that still seem to clash with an effective protection, the European regulation contains elements that, instead, can be appropriately valued in the context of big data, aimed at “technical privacy” and a related preventive approach to data protection risk, based on the accountability and effectiveness of the sanctioning moment.

Alongside the legal issues, the use of big data in the public domain requires paying attention to the profile of their governance: various authorities deal with data (the Agency for Digital Italy, the national authority for data protection and ANAC).

The conclusions will try to highlight the potential of some big data management tools and, consequently, some strategies will be proposed for the management of big data in the public sphere, suitable to protect human rights and fundamental freedoms of the person.


Il diritto all’oblio nei database parlamentari e nella Rete / The Right to Be forgotten in the Parliamentary Databases and on the Internet

Abstract

L’intervento, partendo dalla considerazione che “il trattamento dei dati personali dovrebbe essere al servizio dell’uomo” e che “il diritto alla protezione dei dati personali non è una prerogativa assoluta, ma va considerato alla luce della sua funzione sociale e va contemperato con altri diritti fondamentali, in ossequio al principio di proporzionalità”, si sofferma su alcuni aspetti del bilanciamento tra i diritti dei cittadini di partecipazione alla vita politica, assicurato dalle forme di pubblicità degli atti parlamentari e il diritto alla riservatezza delle persone citate negli stessi nell’era dei Big data. Tale diritto viene considerato tenendo conto delle procedure attualmente seguite dal Senato della Repubblica per rispondere alle istanze di “diritto all’oblio” e alle esigenze poste per la tutela del diritto all’oblio nella rete internet.

“Processing personal data should be carried out for the benefit of man” and “the right to the protection of personal data is not an absolute prerogative, but it should be seen in the light of its social function and it should respect other fundamental rights, in accordance with the principle of proportionality”. Starting from these premises the intervention focuses on some aspects of the balance between the rights of citizens to participate in political life, ensured by the forms of publicity of parliamentary acts and the right to privacy of the people concerned in the age of Big Data. The procedures currently followed by the Senate of the Republic are fulfilling the requests of protection concerning whether the “right to be forgotten” or the “right to be forgotten” on the internet as well.


Intelligenza artificiale e predittività nella risoluzione delle controversie / Artificial Intelligence and Predictive Models for Dispute Resolution

Abstract

Riflessioni sulle applicazioni dell'intelligenza artificiale a supporto delle attività del giudice, con particolare riguardo a metodi e tecnologie che consentono l'analisi dei dati in funzione predittiva.

Some reflections on the applications of artificial intelligence in support of the activities of the judge, with particular attention to methods and technologies that enable predictive data analysis.


La Dichiarazione dei Diritti in Internet e il Regolamento Generale sulla Protezione dei Dati Personali sulle decisioni basate su trattamenti automatizzati: primi passi contro l’abuso del Data Mining? / The Italian Declaration of Rights in the Internet and the European General Data Protection Regulation on Decisions Based on Automated Processing: First Steps Against the “Data Mining Abuse”?

Abstract

Per Data Mining (DM) si indica l’attività di analisi di grandi quantità di dati con lo scopo di rivelare nuova “conoscenza” evidenziando correlazioni nuove, non triviali idonee ad essere implementate in ulteriori processi decisionali. Sebbene non si tratti di una novità, il DM sta acquisendo grande importanza ed attenzione in quanto si presta ad essere impiegato non solo in ogni ambito della conoscenza e della ricerca  ma anche nella modernizzazione delle strategie di organizzaione.

Di fatto la società dell’informazione ha raggiunto un tale livello di realizzazione che la produzione di dati utili all’analisi appare continua e largamente attribuibile a miriadi di fonti coincidenti con quegli stessi soggetti che si servono quotidianamente di ogni tipo di strumento tecnologico. Lo sfruttamento di tali enormi e sempre crescenti quantità di dati per supportare procedure di decisione è al cuore dello sviluppo delle più moderne strategie di organizzazione economica che cercano, così, di essere al passo con le esigenze di un “Data Driven World”.

Ma nonostante l’applicazione entisiastica del DM in ogni campo della conoscenza e dello sviluppo economico, molte critiche sono state mosse in particolare con riferimento all’impiego di tali tecniche da parte dei più potenti enti economici e con riguardo alle conseguenze delle scelte discrezionali che sono necessarie per la raccolta di dati e la stessa conduzione dell’analisi. In questo intervento intendo dare una breve sintesi generale sul concetto e le applicazioni di DM nonché sullo stato delle critiche sollevate in letteratura. In secondo luogo, porrà l’attenzione sulla questione più problematica che riguarda la nozione di “conoscenza” che è in gioco, per mostrare che l’uso comune di tale termine riferito al DM nasconde almeno due diverse definizioni suscettibili di portare a diverse conseguenze giuridiche. Infine, prenderò in considerazione due atti normativi molto diversi tra loro quali la Dichiarazione dei diritti in Internet da un lato e il nuovo Regolamento Generale sulla Protezione dei Dati Personali per cercare di capire se e in che termini vi si possa trovare qualche forma di tutela di fronte ai casi ritenuti più significativi di “abuso di DM”.

Data Mining (DM) is defined as the analytical activity aimed at revealing new “knowledge” from data, highlighting non-trivial patterns useful for further implementation in decision-making processes. Although by no means new, DM techniques are receiving greater attention and relevance nowadays, as they lend themselves to application in almost every field of human knowledge and research as well as in the modernization of business management. As a matter of fact, the development of the information society has reached such a high level of implementation that the production of data potentially useful for analysis is continuous, and largely made by a myriad of sources loosely attributable to the very same subjects who ordinarily use any kind of technological tool. The exploitation of such enormous amount of data for supporting decision making procedures is at the core of the development of more modern business strategies trying to fulfill the requests of the so called “Data Driven World”. But along with the enthusiastic application of DM techniques in every field of research and practical economic development, some criticisms have also been raised focusing in particular on the uses of DM by powerful economic agents, and on the consequences of the inner highly discretionary choices that the analyst or the data collector are bound to make in order to make the DM possible. In this talk, first I briefly give an overview of DM, and of some of the most relevant criticisms raised so far. Then I focus on the most problematic issue of DM, that is the ambiguous use of the term “knowledge”, showing that the general use of this word hides are at least two different meanings each bearing some different consequences in terms of legal compliance.

Than I will take into consideration two very different normative acts, the Italian Declaration of Rights in the Internet (2015) and the European General Data Protection Regulation (2016) in order to evaluate if we can find there a minimal form of legal protection against the most relevant hypotheses of DM “abuse”.


Linkoln, il software per l’estrazione automatica di riferimenti legislativi e giurisprudenziali da testi in lingua italiana / Linkoln, the Italian Solution for the Automatic Identification of Legislative and Case-law References

Abstract

Le citazioni a norme e precedenti rappresentano una delle informazioni di maggiore valore che siano contenute all’interno di un testo legislativo o giurisprudenziale, e risultano di fondamentale importanza nel processo di comprensione di tali testi. Dal momento che esse, tipicamente, non presentano tratti strutturali o di formattazione che le rendano immediatamente distinguibili dal resto del testo e dal momento che sono formulate nelle maniere più svariate, nasce l’esigenza di sviluppare strumenti informatici in grado di identificare automaticamente le citazioni, metterle in evidenza ed agevolare l’accesso agli atti citati tramite l’uso di identificatori standard o collegamenti ipertestuali.

Linkoln, progettato e sviluppato dall’Istituto di Teoria e Tecniche dell’Informazione Giuridica del Consiglio Nazionale delle Ricerche, assolve a questa esigenza: a partire da testi scritti in lingua italiana e attraverso un complesso processo di elaborazione del testo, è in grado di restituire sia l’elenco dei riferimenti legislativi e giurisprudenziali presenti, sia il testo originale annotato con collegamenti ipertestuali in corrispondenza delle citazioni.

Linkoln si affida a una pipeline di servizi che analizzano il testo al fine di identificare, normalizzare ed annotare le potenziali features testuali dei riferimenti (autorità emanante, tipi di documento, data, numero, etc.). L’annotazione di tali elementi basici permette di realizzare ulteriori servizi in grado di individuare pattern testuali via via sempre più complessi, fino ad arrivare all’identificazione di un vero e proprio riferimento legislativo o giurisprudenziale.

Al fine di garantire un alto grado di efficienza del software, i servizi di elaborazione del testo sono stati realizzati con JFlex che, attraverso la definizione di espressioni regolari, regole e stati condizionali, permette di sviluppare automi per l’analisi testuale estremamente efficienti. Linkoln può così essere integrato in un qualunque ambiente Java, sia per l’analisi massiva e off-line di interi corpus di documenti, sia all’interno di servizi web per i quali sia richiesta una certa rapidità di risposta.

Gli identificatori standard nazionali ed internazionali per i riferimenti legislativi e giurisprudenziali attualmente supportati sono urn:nir, Celex, Eli ed Ecli. Per la generazione degli identificatori Linkoln supporta sia una modalità di calcolo autonomo dell’identificatore a partire dai valori normalizzati delle features del riferimento, sia una modalità di interrogazione di repositories autoritativi di riferimenti che forniscano il look-up dell’identificatore corretto a partire dalle features del riferimento.

Con il progetto Linkoln, ITTIG-CNR ha messo a disposizione al pubblico dominio un ambiente efficiente ed affidabile per l’estrazione di riferimenti da testi in lingua italiana. La sua recente adozione da parte del Senato della Repubblica ha mostrato la robustezza del software anche rispetto a testi precedentemente annotati in html o xml. Ulteriori applicazioni calate nel mondo del Web Semantico e dei Linked Open Data sono rese possibili dalla produzione dei riferimenti normalizzati ed associati a standard internazionali, che aumentano le capacità di interoperabilità dei dati e dei documenti legislativi e giurisprudenziali.

References to laws and case-law are among the most important kinds of information that are carried in legislative and juridical texts. Since they do not usually have structural or formatting features that make them immediately distinguishable from the rest of the text and since they are formulated in the most varied ways, there is a demand for tools able to automatically identify citations, mark them up and facilitate the access to the cited documents through the use of standard identifiers and hyperlinks.

Linkoln, designed and developed by the Institute of Legal Information Theory and Techniques of the National Research Council, fulfills this need: given a text written in Italian, it is able to produce both the list of the references to laws and case-law, and the original text enriched with annotations and hyperlinks in correspondence with the found references.

In order to tackle the complexity of automatic reference extraction, Linkoln relies on a pipeline of services which analyze the text in order to identify, normalize and annotate the potential textual features of the references (issuing authority, types of document, date, number, etc.). The annotation of these basic elements allows the creation of further services able to identify more and more complex textual patterns, up to the identification of a whole reference to a law or a case-law.

In order to guarantee a high level of efficiency, the text processing services have been realized with the help of the JFlex library, which, through the definition of regular expressions, rules and states, facilitates the development of extremely efficient automata for text analysis. Being delivered as a Java library itself, Linkoln can be integrated basically into any Java environment, either for an off-line batch analysis of an entire corpus of documents, or to be deployed as a web services for on-the-fly uses.

The national and international standard identifiers for laws and case-law currently supported by Linkoln are urn:nir, Celex, Eli and Ecli. Depending on the specific situation, Linkoln is able to either compose an identifier with the normalized values of the features of the reference, or query an authoritative repository of references that performs the look-up of the identifier starting from the features of the reference.

With the Linkoln project, ITTIG-CNR has made available for the public domain an efficient and reliable tool for the for the automatic detection of references to laws and case-law from Italian texts. Its recent adoption by the Senate of the Republic has shown extremely satisfying performances also with inputs already annotated in html and xml. Further applications in the Semantic Web and Linked Open Data domain are made possible by the production of references associated with standard identifiers, thus increasing the interoperability of the data carried, sometimes hidden, in legislative and juridical documents.


L’intelligenza artificiale e il diritto / AI and Law

Abstract

In a relationship which seems to be apparently of conflict, and in a continuous debate between law and computer science, legal and technical rules arise, stand out, coexist and seem conflicting within the “digitized” society; In this scenario traditional legal categories and liabilities appear more and more mitigated, even confused and with a different dimension due to new roles and functions arisen in the society of these days.

The difficulties typical of a generally universal regulation add up to the ones related to a shared framework of values and clear and effective principles.


Regime giuridico per una profilazione legittima in base al RGPD / Legal Regime for Legitimate Profiling Under the GDPR

Abstract

L’abstract è volto ad analizzare le basi giuridiche che legittimano la profilazione previste dal regolamento generale sulla protezione dati (RGPD), compreso quando la profilazione conduce a processi decisionali automatizzati. Lo scopo è quello di mettere insieme questioni affrontate da altri autori e formulare delle “ipotesi di lavoro” (individuando ad esempio le basi giuridiche più facilmente attuabili per i casi di profilazione). Machine Learning (ML) è un tema decisamente controverso nell’era della protezione dei dati personali. È anche per questo che, tra i vari tipi esistenti, sarà approfondita proprio la profilazione di ML, che è senza dubbio la più frequente, e sarà esaminato come questa è disciplinata all’interno del RGPD.

Il RGPD definisce la profilazione come ogni forma di trattamento automatizzato volto a valutare aspetti personali di una persona fisica. Si fa riferimento a “ogni forma di trattamento automatizzato” piuttosto che a trattamenti completamente automatizzati, che sono invece disciplinati all’articolo 22. Saranno analizzate sia la profilazione volta a prendere decisioni automatizzate ex articolo 22, sia la profilazione come tipologia di trattamento, di cui all’articolo 4(4).

In base all’articolo 22, l’interessato ha il diritto di non essere sottoposto a decisioni basate unicamente su trattamenti automatizzati, compresa la profilazione. Ne deriva che la profilazione non è l’unica forma di trattamento che ricade nell’ambito dell’articolo 22. Ciononostante è difficile immaginare casi in cui decisioni totalmente automatizzate siano state prese senza previa profilazione. Tre sono le basi giuridiche che legittimano questo tipo di profilazione e sono elencate all’articolo 22(2).

All’articolo 4(4) la profilazione è definita come una tecnica di trattamento dati automatizzato, che consiste nell’applicare un profilo a un individuo per prendere decisioni nei suoi riguardi o per analizzare o prevedere preferenze, comportamenti o atteggiamenti. Posto che, in generale, la profilazione non può essa stessa costituire lo scopo del trattamento, ogni qual volta si effettuino trattamenti di profilazione (in realtà ancora prima che si inizi il trattamento), la finalità dovrebbe già essere prefissata e definita. Sarà pertanto determinante individuare la base giuridica che legittima l’ulteriore trattamento dati per il quale è stata utilizzata la profilazione. Tale base legale dovrebbe poter giustificare e supportare che, per raggiungere l’ulteriore scopo del trattamento, sono state utilizzate tecniche di profilazione.

Nonostante la limitazione della finalità fosse già un principio cardine della privacy ben prima dell’entrata in vigore del RGPD, è sempre più idealistico immaginare che le “data companies” riescano a metterlo in pratica.

The abstract is aimed at providing an introduction into the legal grounds for profiling, including where profiling leads to automated decision-making (ADM), under the General Data Protection Regulation (GDPR). The scope is to put together different points touched upon by other authors and collect some working hypothesis (e.g. the more feasible grounds for profiling). Machine Learning (ML) is a highly contentious issue within data protection society. Among the numerous types of profiling, I will focus on ML profiling, which is the most common one, and I will explain how the GDPR assesses profiling.

The GDPR defines profiling as an automated form of processing, aiming at evaluating personal aspects about a natural person. It refers to “any form of automated processing” rather than “solely” automated processing, which is regulated under Article 22. I will refer to and analyse the following two categories: profiling for the purpose of ADM under Article 22 and profiling as a type of processing activity under Article 4(4).

Under Article 22, the data subject has the right not to be subject to a decision based solely on automated processing, including profiling. It follows that profiling is not the only type of processing falling into the scope of Article 22, however in the context of ML it is indeed the one which most commonly comes into play. Profiling in the context of ADM may only be performed relying on three legal grounds, which are listed in article 22(2).

Under article 4(4), profiling means an automatic data processing technique, which consists of applying a profile to an individual in order to make decisions concerning her/him either for analysing or for predicting her/his personal preferences, behaviours and attitudes. Assuming that “general profiling” itself cannot be the processing purpose, every time profiling is used (actually before beginning the processing), the purpose to be achieved should be known. Therefore, the key point is to determine the specific legal ground, which allows further data processing for which profiling will be used. These legal bases should also support that profiling techniques are used to achieve the processing purpose. 

Even before the GDPR, the purpose of data processing to be defined before collection was already a stated principle. However, it still remains quite idealistic for data companies due to practical considerations.


Un algoritmo per aiutare la digitalizzazione dell’ABF / An Algorithm to Help the Digitalization of ABF

Abstract

Nell’ambito delle sue funzioni, l’Unione Europea è focalizzata sulla creazione di un unico spazio dove ciascun cittadino dei vari Stati Membri possa difendere i propri diritti attraverso gli stessi rimedi giurisdizionali. In particolare, l’Unione Europea presta attenzione alla figura del consumatore, prevedendo specifiche procedure.

Con questo scritto, si vuole analizzare, la risoluzione alternativa delle controversie (ADR), e la rispettiva informatizzazione, la risoluzione online delle controversie (ODR). Le relative fonti normative sono da individuare nella Direttiva sull’ADR per i consumatori, Direttiva 2013/11/UE del Parlamento Europeo e del Consiglio del 21 maggio 2013, ed il Regolamento sull’ODR per i consumatori, Regolamento 524/2013 del Parlamento Europeo e del Consiglio del 21 maggio 2013.

In Italia, questi atti normativi hanno trovato diverse applicazioni, in particolare con l’istituzione dell’Arbitrato Bancario Finanziario (ABF), ovvero di una ADR presso la Banca d’Italia, in ordine all’art. 128 bis del Testo Unico Bancario (TUB). Lo scopo di quest’autorità è quello di risolvere controversie fra consumatori (persone fisiche e giuridiche) e gli intermediari finanziari, in tema di transazioni e servizi bancari e/o finanziari.

Al momento non è prevista una corrispettiva ODR, ma può essere sviluppata analizzando e risolvendo specifiche situazioni. In questo paper è formalizzato il caso specifico avente ad oggetto lo strumento della cessione del quinto, con specifico riferimento alla richiesta di rimborso delle somme versate in caso di cessione anticipata del contratto.

Infatti, in forza dell’art.125 sexies del TUB, il consumatore ha diritto al rimborso anticipato in qualsiasi momento dell'importo dovuto al finanziatore ed in tal caso ha diritto al rimborso proporzionale di interessi e costi dovuti. Proprio sulla riduzione del costo totale del credito l'ABF è continuamente chiamata a decidere: i consumatori si rivolgono a tale ADR proprio per vedersi riconosciuto il congruo e proporzionale rimborso di quelle che, in dottrina e giurisprudenza, sono concordemente definiti costi c.d. recurring. Scendendo nel dettaglio, l'Arbitro viene chiamato a quantificare tali rimborsi e lo fa applicando un criterio emergente dallo stesso art. 125 sexies TUB ovverosia quello del c.d. pro rata temporis che valorizza il momento in cui si è proceduto all'estinzione.

È possibile verificare come la maggior parte delle decisioni relative a questo caso possano essere risolte anche attraverso un semplice calcolo matematico, facilmente implementabile in un software attraverso un algoritmo, raggiungendo in questo modo gli obiettivi dell'Unione Europea di rispettare i principi di efficienza, trasparenza ed economia a beneficio del consumatore, in particolare sfruttando il ‘mondo online’.

In its role, the European Union focuses on creating a single place where all citizens of each Member State can defend their rights through the same judicial remedy. In particular, the EU gives considerable attention to the consumer and has therefore established specific procedures for the consumer.

With this paper, the author wants introduce one of this procedures, the alternative dispute resolution (ADR), and its computerization, the on-line dispute resolution (ODR). The reference legislation is the Directive on consumer ADR, Directive 2013/11/EU of the European Parliament and of the Council of 21 May 2013, and the Regulation on consumer ODR, Regulation (EU) No 524/2013 of the European Parliament and of the Council of 21 May 2013.

In Italy this legislation found application in many cases, in particular with the implementation of the Banking and Financial Ombudsman (ABF, Arbitrato Bancario Finanziario), an ADR set up at the Bank of Italy, pursuant to Article 128 bis of Consolidated Law on Banking, introduced by the Investor Protection Law (Law 262/2005 or TUB, Testo Unico Bancario). The aim of this authority is to resolve dispute between consumers (natural persons but also legal persons) and financial intermediaries, concerning banking and financial transactions and services.

At the present time it is not expected a respective ODR method, but it can be developed thanks to the realisation of individuals sectors to resolve specific causes. In this paper it is expressed as a formula the cause concerning banking and financing operations on the granting of cessione del quinto (an employee loan repayment unique to Italy), with specific reference to the request for reimbursement of expenses incurred by the applicants and not accrued in the event of early termination of the relationship.

In accordance with the art. 125 sexies of the Consolidated Law on Banking, the consumer has the right to early repayment at any time of the amount due to the lender and in this case is entitled to proportional reimbursement of interest and costs due. It is precisely the reduction in the total cost of credit that the ABF is continually called upon to decide: consumers turn to this ADR precisely in order to have the congruous and proportional reimbursement of what, according to doctrine and jurisprudence, are commonly defined as ‘recurring’ costs recognised. In particular, the Ombudsman is called to quantify these reimbursements and does so by applying a criterion emerging from the same art. 125 sexies of the Consolidated Law on Banking, namely that pro rata temporis which values the time when the extinction took place.

It is possible to verify how most of the decisions concerning this case can be resolved also through a simple mathematical calculation, easily implementable in a software through an algorithm, achieving in this way the objectives of the European Union to respect principles of efficiency, transparency and economy to the benefit of the consumer, in particular by ‘the online world’.


EU Vocabularies – Facilitating the Linking of Legal Data

Abstract

The Publications Office of the European Union is an interinstitutional office entrusted with the task of publishing EU law and publications in the 24 official EU languages. To face the challenge of publishing multilingual legal data, the Publications Office makes use of multilingual controlled vocabularies such as the EuroVoc thesaurus and the authority tables (ranging from lists with language codes to the names of the EU institutions and bodies). EuroVoc and many other controlled vocabularies produced by the EU Institutions are available as open data on the new EU Vocabularies (https://publications.europa.eu/en/web/eu-vocabularies/) website, the EU Open Data Portal (https://data.europa.eu/euodp/data/dataset?tags=controlled+vocabulary) and through the SPARQL endpoint (https://data.europa.eu/euodp/data/dataset/sparql-cellar-of-the-publications-office) of the Publications Office.

EuroVoc is a multilingual multidisciplinary thesaurus covering the activity fields of the EU institutions. It is expressed using semantic web standards recommended by the W3C such as SKOS (https://www.w3.org/2004/02/skos/). Besides modelling flexibility and reasoning capabilities, semantic technologies also facilitate the alignment of EuroVoc with other thesauri and controlled vocabularies, such as national and domain specific vocabularies.

The presentation will focus on the challenges faced during the alignment exercise with the national Legilux vocabulary of the Luxembourgish government and EuroVoc as well on the opportunities that arise from this linking of vocabularies such as improved interoperability and information retrieval. The aim of the alignment is to build semantic interoperability based on thematic classification between national legislations and EU legislation, but also from Member State legislation to Member State legislation. The objective of the project is to test, improve and validate methodology and tools to build and maintain those alignments. The methodology is based on the fact that EU legislation is annotated with thematic subjects based on EuroVoc, and that EU legislation is transposed in national legislation texts which are annotated with a national thematic vocabulary. The presentation will introduce the use cases of the establishment of semantic relationships between data items within different data sources by using the alignment tool SILK (http://silkframework.org/) and it will propose a methodology to build the alignments from a statistical, linguistic and manual approach.

The presentation will also demonstrate the advantages of using the open source vocabulary management tool VocBench (https://joinup.ec.europa.eu/solution/vocbench3/about) that includes a built-in alignment validation function. Designed to meet the needs of the semantic web, the tool manages multilingual controlled vocabularies such as thesauri, authority lists and glossaries. As a collaborative tool it also allows users to maintain, validate and publish content through its flexible group management environment.


IT Law Modelling: Methodologies, Architecture and Ontologies: Critical Legal Positivism Applied in the Domain of IT Law

Abstract

This study participates in the doctrinal discussion on IT law, but from a somewhat unconventional angle of applied legal theory. The objective of this study is to establish the possibility of cognition of the subject matter of IT law and to formulate its constitutive categories and descriptive ontologies. The chosen approach is multidisciplinary, with a legal standpoint relying on Tuori´s critical positivist theory of law´s levels systems, and a knowledge engineering standpoint borrowing languages and methods from software and ontological engineering to represent knowledge in an explicit, shareable and maintainable manner.

The use case considered consists in the provision of models and reference material to cover the legal knowledge needs of IT practitioners. For IT professionals, access to knowledge on IT Law remains problematic and keeping abreast of the developments on compliance requirements is a daunting task. Still, they need to understand their legal position in order to anticipate possible disputes. We postulate that models and ontologies can bridge the knowledge gap and serve as lingua franca between the legal and the IT profession.

A model is a representation of a reality, an abstraction, a simplification, a depiction. Modelling law can be a necessity for a number of reasons: legal drafting, analysis of court cases, development of computer programs implementing/supporting law, teaching law. This paper discusses the pros and cons of using semi-formal models for representing the law with didactic purposes. Languages and methods for software and ontological engineering make explicit an idea that has been traditionally part of legal thinking: the idea that legal concepts have a structure and are linked one to another.

This paper contends that the intermediation of semi-formal models and their derived diagrams and schemas are helpful towards improving the legal literacy of IT professionals. The use case considered will be the provision of a model of the EU legal regime on cookies and associated reference material meant to be employed by human users to learn this subject. Following the principle of Biggs´constructive alignment and Bloom´s taxonomy of learning objectives, the learning outcome for the audience of IT professionals will be to know, understand and apply the legal definitions, rules, rights, obligations and legal conditions of lawfulness.

The objectives of this paper is to assess the effectiveness of IT law modelling to cover the legal knowledge needs of IT practitioners; the modelling formalism for the purpose of modelling IT law for this use case will be Unified Modelling Language. Consequently, the research question addressed in this paper can be formulated as follows:

Is Unified Modelling Language suitable to teach IT law to an audience of non-legal experts?


Semantic Finlex: Finnish Legislation and Case Law As a Linked Open Data Service

Abstract

Everybody is expected to know and obey the law in today's society. Governments therefore publish legislation and case law widely in print and on the web. Such legal information is provided for human consumption, but the information is usually not available as data for algorithmic analysis and applications to use. However, this would be beneficial in many use cases, such as building more intelligent juridical online services and conducting research into legislation and legal practice. To address these needs, this paper presents Semantic Finlex, a national in-use data resource and system for publishing Finnish legislation and related case law as a Linked Open Data service with applications. The system transforms and interlinks on a regular basis data from the legacy legal database Finlex of the Ministry of Justice into Linked Open Data, based on the new European standards ECLI and ELI. The data is hosted on a “7-star” SPARQL endpoint with a variety of related services available that ease data re-use. Rich Internet Applications using only SPARQL for data access are presented as first application demonstrators of the data service.

 

Governments provide publicly available legal information on the web usually in the form of HTML or PDF documents targeted to human readers. In Finland, for example, legislation and case law are published as HTML documents in the Finlex Data Bank (http://www.finlex.fi), a publicly available online service since 1997, maintained by the Ministry of Justice. However, Finlex does not provide publicly available machine-readable legal information as open data, on top of which services and analyses can be built by the ministry or third party vendors.

 

Authorities in Europe strive to improve the semantic interoperability between EU and Member State legal systems, as the methods in use now for storing and displaying legal documents differ among countries. Therefore, the Council of the European Union has invited the introduction of ELI (European Legislation Identifier) and ECLI (European Case Law Identifier) standards that define common identifier and metadata models for legislative and case law documents by applying Linked Data principles.

 

Semantic Finlex, a national Linked Open Data Service for Finnish legislation and case law addresses these problems and trends based on new European and Semantic Web standards. The service hosts and publishes a central part of the Finnish legislation along with judgments of the Supreme Court and the Supreme Administrative Court. All of the datasets are automatically updated regularly.


Serving the Publics: On Metadata Portability

Abstract

Legal systems are an important element of a broader public discourse in all jurisdictions. In the global, interconnected, high-tech, and sometimes over-heated environment of today, legal institutions are an important bulwark against rumor and opportunistic misrepresentation. Fulfillment of that role turns not only on accessibility, but on ready digestibility, both in the sense of plain language, but also in the sense of plain mechanics. Put simply, the easier it is to add citations and links to legal sources to a newspaper article, a blog post, or a polemic screed, the more likely it is that will happen; and exposure is the mother of expectation amongst the reading public.

While neural networks and big data capture headlines and the attention of players keen to preserve or secure position in a rapidly changing environment, the fact remains that the core value of law lies in its structure, and that releasing its value to machines as well as humans depends on metadata—which for both “audiences” today means metadata machine-readable form. This is an area in which the law once excelled, but now lags behind the sciences and the general publishing trade (even the advertising industry). Some jurisdictions have made great strides, but national and federal efforts are stovepiped, in much the same way as was once true of character encodings.

This presentation and paper will review the status quo in online provision of metadata in selected legal archives, outline the requirements for machine-assisted authoring (reference management) drawing on resources from multiple jurisdictions, and set forth first steps that archive maintainers can take to help walk the law, in all its instances, toward a smoother user experience.


Towards a Linked Open Data Cloud of Language Resources in the Legal Domain

Abstract

The work described in this paper is framed in the H2020 Lynx project, which is aimed at developing legal compliance services across different languages and legislations, based on a legal knowledge graph (LKG). The legal knowledge graph integrates and links heterogeneous compliance data sources including legislation, case law, standards and other private contracts.

In order to stablish the foundations of this legal knowledge graph, a sound set of language resources needs to be generated. Language resources are understood as pieces of structured data in machine-readable form such as corpora, terminologies, glossaries, lexicons or dictionaries. Within Lynx project, language resources are required to annotate and classify legal documents, train machine translation tools, test natural language processing algorithms and related activities.

The first stage of the process described in this paper comprises the identification of existing legal language resources. Some examples are Jurivoc (https://www.bger.ch/ext/jurivoc/live/de/jurivoc/Jurivoc.jsp?interfaceLanguage=german), a juridical thesaurus for Swiss regulations; the UNESCO thesaurus (http://skos.um.es/unescothes/?l=en), which contains terms from various fields including the legal domain; or the STW thesaurus (http://zbw.eu/stw/version/latest/about), covering the economy domain. These resources are available in different formats such as CSV, XML or PDF. Since the value of such assets increases when interlinked, the next step is the conversion into RDF, which is a good choice to represent structured information and metadata, and to easily create links between resources. As a result of this linking process, a Linguistic Legal Linked Open Data (LLLOD) cloud, core of the Legal Knowledge Graph, is being generated as part of the Linguistic Linked Open Data (LLOD) cloud, in turn a subset of the Linked Open Data cloud.

Part of the conversion work consists in selecting the vocabularies and properties to be applied to the datasets, heavily relying on SKOS vocabularies, that represents concepts and hierarchical relations between them, and the Ontolex vocabulary, that is used to model linguistic information, have been considered for this purpose.

More language resources are to be identified and converted into RDF and eventually, specific legal corpora provided by Lynx project will be used to create new language resources thanks to the implementation of automatic term extraction techniques. Both sets of resources, the one identified and the one created, will be linked to contribute to the enrichment of the Semantic Web in the legal domain.


DaPIS: A Machine-readable Data Protection Icon Set

Abstract

Privacy policies are known to be impenetrable, lengthy, tedious texts that are hardly read and poorly understood. Therefore, the new EU legal framework for data protection, the General Data Protection Regulation (GDPR), introduces provisions to enhance information transparency and suggests icons as visual means to clarify data practices.

Notwithstanding the many benefits in terms of e.g. comprehension that legal visualizations demonstrably provide, visual communication can take many different shapes. The scientific debate around graphical symbols for legal concepts is still in its infancy. Both the creation and consequent evaluation of icons depicting abstract or unfamiliar concepts represent a challenge. Moreover, precision of representation can support the individuals’ sense-making, but at the expense of simplicity and usability.

We present DaPIS, the Data Protection Icon Set that we created and evaluated through human-centred methods drawn from the emerging discipline of legal design. Firstly, we have organized rounds of participatory design sprints where designers and lawyers collaborated side by side. Then, we ran some user studies to empirically determine strengths and weaknesses of the icon set as communicative means for the legal sphere.

 The icon set is modelled on PrOnto, an ontological representation of the GDPR, and is organized around its core modules: personal data, roles and agents, processing operations, processing purposes, legal bases, and data subjects’ rights. In combination with the description of a privacy policy in the legal standard XML Akoma Ntoso, such an approach makes the icons machine-readable and semi-automatically retrievable. Icons can thus serve as information markers in lengthy privacy statements and support the navigation of the text by the reader. In this way, we aim to map and connect different representations of legal information to enhance its comprehensibility: the lawyer-readable, the machine-readable, and the human-readable.

More information about the research can be found at: http://gdprbydesign.cirsfid.unibo.it.


From Parliamentary Open Data Visualization Toward Gamified Civic Engagement

Abstract

In this paper we define a set of guidelines for the design of a gamified eDemocracy and eParticipation tool based on Open Governmental Data . Firstly, we present the data visualization method that we have used to make Open Governmental Data more understandable for citizens. Secondly, we describe the technologies and the architecture of the mobile application that we have used to collect the data used by the graphs that we created. Thirdly, after having enumerated the advantages to develop a gamified application rather than a mere graphical visualization of data, we propose a set of gamification guidelines that are useful to improve civic engagement and we exemplify a possible implementation of these. Lastly, we pinpoint some studies that could be performed through the testing of a future implementation of the application with real users and through the analysis of the test outcome.

Data visualization techniques are very useful to represent complex data in a human-readable way. In the simplest scenario, a bill is proposed by some deputies or by some senators, and then it passes through the Chamber of Deputies to the Senate, and vice versa, until the same text is approved or rejected by both chambers. In specific moments of the enacting process, bills could be examined by various commissions that can approve or that can reject them. All these steps are transparent and properly documented.

During the Code4Italy@Montecitorio hackathon, we have developed a set of graphs that simplify the Italian legislative process, the main topics of the bills enacted in Italy in the last years, and the amount of rejected, approved, stalling and yet-to-discuss bills in specific legislations. After having created a website to host all the graphs that we designed, we implemented a mobile application for Android devices. We investigated the methodologies to move from a mere data visualization tool toward an eParticipation application to improve democratic deliberation among users and users’ civic engagement.

Gamification is the “use of game-design elements in non-gamig context” (Deterding, Sicart, Nacke, O'Hara, Dixon, 2011). Recent researches have shown that the use of basic core concepts of game design, more specifically game dynamics and game mechanics in activities that are not games (Mekler, Brühlmann, Opwis, Tuch, 2013), improves the user experience. Moreover, users have fun and are satisfied, even if the activity is complex or not “natively fun”. Gamification revolves around intrinsic and extrinsic motivations of human beings. Extrinsic motivations are related to the need of human beings to receive rewards when they perform specific actions. Intrinsic motivations are related to innate needs of human beings, like autonomy, competency and relatedness, as stated by the Self Determination Theory (Ryan & Deci, 2000), hereinafter SDT. By matching the needs of eParticipation and eDemocracy tools with the gamification components that are suitable to satisfy all the human beings’ needs stated by the SDT, we have found a set of gamification guidelines that are suitable to create gamified eParticipation environments that can improve citizens’ civic engagement.


From Transparency to Terms, and Back Again: A Legal Design Approach

Abstract

The information paradigm is the backbone of two important areas of European law: consumer and data protection. It postulates that enhancing the transparency about the fundamental elements of a transaction or a data processing better protects the rights and interests of the individuals involved. However, to read online Terms of Service (ToS) or privacy policy is not always a widespread practice among users (see, for example, Eurobarometer Special 447 on digital platforms, 2016).

Several cognitive studies explain the reasons of such an attitude: ToS and privacy policy are usually long, badly written, and full of legal jargon; at the same time, individuals often suffer from several forms of bias and degrees of illiteracy. Therefore, users may find quite hard to read and fully comprehend the basic conditions of the agreement and/or its legal consequences (such as the identity of the counterparty, the rights and remedies available, the allocation of responsibilities, the fairness of the clauses). Other users, instead, may decide to “click” without reading because of the impossibility to negotiate those terms. Therefore, the current state of implementation of the information duties and bounded rationality are the two main reasons why the information paradigm is failing its policy goal.

My research hypothesis is that smart disclosure mechanisms may help to fix this failure. The paper will show the theoretical background and the preliminary results of the “Awareness by design app”, developed within the project “The Internet of Platforms: an empirical study on private ordering and consumer protection in the sharing economy”. The methodology for this research is based on legal design, an empirical approach that combines behavioural studies, technology and design to solve legal problems. The proposed app will enhance the information users receive from and about the digital platform, thanks to the combination of two different behavioural strategies: labels and ratings. On the one hand, the app will inform the user about the key conditions of the platform’s terms and privacy policies through icons and short texts; on the other hand, the app will provide a score system to rate the quality of the above mentioned legal documents. The research is based on the empirical analysis of a representative set of digital platforms operating at the European level in three different sectors (transportation, accommodation, food) and the app is being tested with real users in the Brussels-Capital Region.


Legal Design: Applying the Design-driven Approach to Legal Services

Abstract

Legal Design is the conjunction between law and design for re-thinking the legal practice: using the design-driven approach to make the legal services user-centered and delivered more efficiently (Margaret Hagan; 2017).

By implementing the Design tools and ideas, it could mean rendering the legal world more friendly and user-centered. Since Law is a social career –and it could be considered as a final service- all the methodologies from Design Thinking could be applied. According to Brown, T., & Katz, B. (2009), Design Thinking makes the tools and skills designers have learned over time available to non-designers so they can apply them to a new range of problems. This perspective introduces a main change: you can make decisions based on analyzing future users and not only on historical data or instinct. “Design Thinking is a human-centered approach to innovation that draws from the designer’s toolkit to integrate the needs of people, the possibilities of technology, and the requirements for business success.”  (Tim Brown; CEO OF IDEO; 2017)

As a result, Legal Design is an attempt to innovate in the legal field and transform the practice to answer the demands of the 21st century. Applying the design methods to the legal world can mean making Law more approachable and changing the view non-layers have of it today. Moreover, it can help develop easier and simpler legal processes that will affect our society in a direct way by making it more transparent and democratic.

The aim of this paper is to understand what is Legal Design and its scope. By analyzing case studies, it was possible to find common areas explored by start-ups, companies, and governments. The result is the following mind map that explains the implementation of Design in Law. It shows the clusters found during the desk research and it explains how each field of design collaborates in the innovation of the legal process. It is also necessary to evidence that some topics are more recurrent than others. As well, there are some countries -and regions- where Legal Design is more developed.

As a secondary objective, it is important to reflect about the possible transformations that Law -and Design- might experience. Designers can have a new role inside of the legal world: there are some cases where they work inside of law firms; but there are some designers who started thinking in a deeper level, related to democracy, by working on a government stratum.

After understanding Legal Design, it is possible to claim that implementing Design Methodologies in the Legal World could mean changing the vision that Non-Lawyers have of Law. So, this could mean creating clear processes, including bureaucratic ones. It can help to develop easy and simple legal process that will affects in a direct way our societies making it more transparent. Lastly, in this way, citizens would be empowered by reading and interpreting Laws and Legal topics.


Visualizing Law: Experiences and Perspectives of Legal Visual Analytics

Abstract

Social sciences are undergoing an era of deep change. The data deluge and the availability of advanced computational heuristics are providing researchers with innovative ways to delve into social complexity. The same applies to the legal world, where techniques spanning from NLP to machine learning and network analysis are opening up new epistemological and methodological prospects to both research and practice. The presentation will focus on the legal applications of visual analytics. After a brief introduction to the rise and the challenges of what we define as Visual Legal Analytics, we present three experimental platforms for visual analytics discussing their objectives, implications, and perspectives.


Anonymization Service for Finnish Case Law: Opening Data Without Sacrificing Data Protection and Privacy of Citizens

Abstract

The Finnish public sector produces vast amounts of valuable data that concern private citizens of Finland. The public sector highlights that there is a growing need for anonymization but also a number of challenges. The current state of anonymization tools is that each actor of the Finnish public sector takes care of the anonymization of their own documents. The Finnish court case data, for example, could be useful in governmental decision making and research if opened for the public. However, due to issues of data protection and privacy it is not possible to share the data openly. This challenge is especially urgent now that the General Data Protection Regulation (GDPR) came into effect on May 25, 2018. Prior to the GDPR, the court cases were available and could be obtained in print for a request from courts, but their publication on the web was restricted. Now the court cases need to be sealed from the public, unless identifying details of persons and companies is removed from them.

Anonymization is the process that removes explicitly or implicitly identifying details of persons and companies from text. Whereas the procedure of substituting identifiable information with neutral names, i.e. pseudonyms (such as ”person A”), is called pseudonymization. In order to release court case documents to the public it is required that the documents are at least pseudonymized. The proceedings of the decision making in courts have been highly interesting for the public but few decisions have been anonymized and made public. The anonymization of the decisions is expensive because it requires more human resources. In this paper, we propose an automatic pseudonymization method and tool for the court decisions.  Currently, we are in the process of creating a semi-automatic service for Finnish and Swedish languages. The application pseudonymizes  documents and the user can fix and approve the results before publication. The success of the tool depends on the applicability of the user interface in addition to the precision and recall of the language technology tools. In order to measure the functionality of the service, its performance will be compared with the manual pseudonymization estimates. The number of mistakes and the time it takes to fix the mistakes will decide the usefulness of the service in the case of anonymizing Finnish court decision materials.


ECLI Implementation in Cyprus: Organizing Legal Information in a Small, Mixed Jurisdiction

Abstract

A former British colony and an EU Member State since 2004, the Republic of Cyprus is a mixed legal system, in comparative-law terms. English law dominates most of private law, including contracts, commercial and company law, criminal law and procedural law across the board and the judiciary has a common-law mentality. At the same time, Cypriot family law and public law is strongly oriented towards the Continental legal tradition. The effective bilingualism of the legal system, with Greek as the official language and the language of the population and English as the original language of many legal sources and much of legal practice, adds to the mix.

For a long time, access to legal information was restricted to a privileged few (in a system moreover where case law constitutes a source of law), but Cyprus has been recently moving dynamically to open-access legal information.

The paper discusses the challenges of implementing ECLI in Cyprus – the first common-law or mixed jurisdiction to do so, with the help of an EU-funded project, CyECLI, which concludes in late 2018. After presenting the state of play with regard to access and organization of legal information prior to the implementation process, the paper examines the policy and technical choices made in ECLI implementation in Cyprus and considers lessons to be drawn for other common-law jurisdictions. The paper also considers the relationship between ECLI and the neutral citation systems of common-law jurisdictions, the role of court reports (another common-law institution) in an era of online registries and automated access to information, and the potential impact of ECLI 2.0 for Cyprus.


How Much Appetite for Case Law Open Data in France?

Abstract

The French law of October 7, 2016 “for a digital republic” provides with an open data of all court decisions. Only decisions of supreme courts and those selected by courts of appeal have been distributed online for now. However the related application decree has not been issued yet. The volume of decisions at stake, the backlog of digitising needs of the judiciary and governance issues were already strong constraints. Since the law was passed, the increasing sensitivity of the public to personal-data protection and concerns among the judiciary about possibly nasty side-effects by open-data-based, AI-enabled, so-called “predictive” justice, have added to the challenges. On the other hand, the government recently renewed its commitments to the Open Government Partnership, which includes the opening of case-law. All these factors weigh on the actual scope and timing of the project.

My intervention will be therefore structured as follows :

  • I will draw up an inventory of the current judicial and legislative open data offering in France
  • guess the content of the decree dedicated to case-law open-data, or analyse it if it is published by the time of the conference
  • re-assess the potential of re-use of case-law and legislative data given likely upcoming restrictions on case-law open-data policy

Open Data and the ECLI at the Italian Constitutional Court

Abstract

The Court’s most important task is the review of the constitutionality of the laws and other legislative acts of the State, as well as of the laws enacted by territorial communities such as the Regions and the Autonomous Provinces of Trento and Bolzano. The decisions adopted by the Court are judgements (sentenze) and orders (ordinanze). From 1956 to date, the Court has published approximately 20,000 decisions, all of which are available on the Court’s official website. In 2012, the Court decided to make the documents held in its archives available for re-use and for any purpose, starting with its decisions. The decisions were initially published in an open format (XML), which was already being used for publication on the Italian Official Journal and, later, in JSON and CSV formats. The data contained in the digitized document are ready for re-use through an open license (Creative Commons), which grants broad rights to re-use and guarantees the non-alteration of the document and acknowledgement of the source.

In 2010, the EU Council Conclusions established the European Case Law Identifier (ECLI). The univocal identifier (the ECLI) can be used for the decisions issued by all courts within the European Union and other European organizations (of the Council of Europe and of the European Union) to make European law and Member states’ case law available to national judges, legal professionals and citizens. At the end of 2016, the Constitutional Court adopted the ECLI: all of its decisions published since 1956 (which, as mentioned above, amount to approximately 20,000) are provided with an identifier having the following format: ECLI:IT:COST: YEAR of the decision: NUMBER of the decision.

In addition, compulsory metadata are available (among which, creator, coverage, date, language). The English translation of the decision, when issued, is also made available via the ECLI search engine.   


Policy Issues in the Publication of Case Law - An Australian Perspective

Abstract

The two main sources of law in Australia are case law (common law or judge-made law) and legislation (Acts – created by members of Parliament).

This leads us to examine policy issues in the publication of case law by comparing publication of these main two sources of law in Australia. The current position is that publication of case law (reports) is available via open access (unauthorised versions) and subscriber access (some unauthorised and authorised versions). The vast majority of legislation (authorised versions) is available via open access (that is, freely available).

In Australian courts there is a convention that the authorised report of a judgment is the version that must be cited in preference to other versions (such as unauthorised or unreported).

This convention has been articulated by practice direction in most jurisdictions. In the remaining jurisdictions, it remains a matter of convention only.

Authorised law reports are published by legal publishers and are only available by subscription. Subscription to authorised reports is expensive and increasingly difficult to justify.

Contrasting the case law experience in Australia with that of legislation makes obvious the difference in access and availability to the two main sources of law in Australia.

Both sources of law are equally important in the administration of justice. There are no cost barriers to accessing authorised versions of legislation in most Australian jurisdictions, yet the cost barriers remain in relation to accessing authorised reports.

Accuracy and credibility issues are overcome when the publication of the freely available online legal content in the form of case law is by the 'source' being the court and the judiciary that made the decision; and where an organised system of citation is embedded. Online platforms that provide a certification system verifying the accuracy of the reproduction of the original decision is another safeguard.

Convention and practice directions continue to require reference to authorised reports however it is time for the courts to respond to advances in technology and the new socio-legal reality that has resulted by embracing the fact that freely available online legal content is the main legal resource used by most legal practitioners and that it is consistent with the ideal of equity in access to the two primary sources of law in Australia – case law and legislation.


Brexit: Meeting the Legal Information Challenge

Abstract

The United Kingdom will leave the European Union on the 29th March 2019. The European Union (Withdrawal) Act provides for a fully functioning domestic statute book when the UK leaves the EU. To avoid large gaps appearing in the statute book, the Act converts existing EU law as it applies immediately before exit day into UK domestic law. It also preserves UK laws that have been made to implement EU obligations. Parliament and, where appropriate, devolved institutions, will then be able to make changes to these laws post-exit.

To meet the demands of legal certainty and transparency, this new body of law needs to be published and provided to the public, ideally with a statement of the text as it applies in the UK. This is a significant challenge for the UK Government’s official source of newly made and revised legislation – legislation.gov.uk.

At first glance, the proposal that EU law will continue to have effect in the UK post-exit might appear to make it easy for a UK citizen to find out what laws they are subject to once the UK has left. After all, EU legislation is all available, for free and in a variety of formats, on EUR-Lex.

However, while it is true that on exit day the law of the EU and the retained EU law in the UK may appear to be identical, the UK is no longer subject to the former, and the EU has no control over the latter. The two versions of EU law, from exit day, will start to diverge. This presents a unique challenge to legislation.gov.uk.

Many EU countries have integrated EU legislation into their domestic online legislation platforms, but this is the first time that any country has attempted to fully domesticate it, with all of the complexity which that entails. The legal information challenges for legislation.gov.uk can broadly be split into the following categories:

  • Provenance – how to source the legislation for retention, and provide evidence that it is accurate.
  • Publication – the Acquis Communautaire will form part of the UK statute book. It will therefore need to be captured and transformed, with no loss of fidelity, into a format that allows it to be maintained alongside existing UK legislation.
  • Interoperability – the retained EU legislation will be UK domestic law, which means the two categories of legislation will interact directly in ways not undertaken previously.
  • Presentational – how to ensure that this new body of law is published in a way that provides maximum legal certainty and confidence for legislation users.

This presentation sets out the work being undertaken by The National Archives to meet the legal information challenges and opportunities posed by the UK’s exit from the European Union.


Building Commentary Collections in the Free Law Context

Abstract

When the Free Access to Law movement started, the vision to make the law available online for free was aspirational. While comprehensive coverage of primary law has proven to be a moving target, the growing body of publicly available information and ongoing improvements in platforms has been a success in many countries. This success has shown that while access to primary law is a precondition to access to legal information, it is not sufficient for true access by the legal community or the public. This shows the necessity for expanding access to other content types as the logical next step for free law online.

At CanLII, for the last year we have been working on improving access to commentary resources that will support this mission in Canada. This has been facilitated by the combined developments of the technological tools that allows it and an emerging social context that supports it. The biggest constraint we initially faced was having the technology in place to support publishing a large volume of third party content. Once that was in place we could start to explore possible ways to find content to add to the CanLII collection. This session will discuss the different possibilities we explored developing CanLII’s commentary, and the logistics as we encountered them. This will include information about the technological solution, experiences we had soliciting content, and how we developed a plan to fund this development without an increase in core funding.


Digitization of U.S. Congressional Documents

Abstract

We will present a case study on the theoretical and practical concerns that formed the approach at Rutgers to the digitization of a large collection of print government documents. Since 2005, the library has been engaged in the destructive digitization of its collection of U.S. congressional committee hearings and prints. Using standards negotiated with the U.S. Government Printing Office, the congressional documents were withdrawn from our collection and have been getting scanned for publication on the Internet. The project is getting into its later phases, and currently has over 18,000 documents available online.

The overall aims of the project were to preserve and publish the documents, and to do so in as cost effective manner as possible without sacrificing usefulness and readability. We will talk about how we pursued those goals, including in this discussion the following topics:

  • Imaging standards for publication and for archiving
  • Metadata standards
  • Archival standards
  • OCR processing
  • Quality control
  • Labor costs
  • Equipment
  • Workflows
  • Disposal of scanned material

We will also discuss search engine and interface issues. Finally, some experiments in automating quality control, and other further efficiencies will be discussed. In the end, the processes and techniques used allowed Rutgers to conduct this project at a cost that could be handled within our normal operating budget, eliminating the need for the effort and oversight that comes with finding and maintaining outside grants. Some of these savings are peculiar to Rutgers, and some peculiar to the material involved. However, the overall approach may be adaptable to many types of materials and many other kinds of institutions. The possibilities of employing these techniques in other contexts will be an important part of the discussion.


Foundations of the Common Law

Abstract

This presentation will explain the purpose of and progress on a project entitled ‘Foundations of the Common Law Library’. Its scope is the first seven centuries of the common law before the modern era from 1215 (Magna Carta) to 1914 (WWI), and it will include legal resources from up to 70 common law jurisdictions. The project, under development by the Australasian Legal Information Institute (AustLII), has obtained funding of AUD $1.1 million. The Library, to be launched in 2019, will be a free access research facility planned to contain at least half a million full-text pre-1914 cases, legislation, books and treaties, searchable from one location.  The project involves cooperation established over more than a decade with ten other free access Legal Information Institutes (LIIs) across most of the common law world, and a growing number of law libraries. The current version of the Library, which makes searchable approximately one quarter million pre-1914 documents, will be demonstrated, and future directions of the project explained. The Foundations Library will be located on CommonLII (Commonwealth Legal Information Institute).

The theoretical basis of the project is the premise that the courts (and other legal institutions) of common law jurisdictions were more interdependent than is commonly thought. This project will provide the research infrastructure to allow this and other important theses to be explored by researchers on a much larger scale. What was the nature of the common law in the colonies? How did legal variation (innovation) spread from place to place? How much was the law (and its variation) influenced by the circumstances of its location? Did colonial variations return to have an influence on the common law in England itself?  There is a growing body of work in legal history which seeks to move beyond a single jurisdiction  (or direct comparisons) to trace the ways laws moved around the British Empire (as it was then called), particularly the ‘settler’ colonies. This project will support this new trend in legal history. The presentation will explain why such a Library is needed.


The State of Free Online Access to Nigerian Public Legal Information

Abstract

This paper aims to assess the current state of free access to the primary sources of Nigerian law on the World Wide Web, provided by the government and non-state actors. It presents a historical perspective of the publication of laws in Nigeria (one of Africa’s leading economies) and examines the adequacy of its legal framework for access to public legal information in this information and communications technology (ICT) age. The paper reveals that neither the Nigerian federal government nor any of its thirty-six States has any official online legal information database and that the prevalent practice by these governments is the publication of print legal information as a revenue-generating business. It further finds that non-state actors provide access to a majority of the available free online legal information resources, but they are grossly inadequate because they are neither up-to-date, comprehensive, nor accurate reproductions. The paper argues that the lack of political will to actualise the basic tenets of democracy (especially, transparency, the rule of law, and public participation in governance) is the preeminent factor responsible for the extremely poor state of public access to Nigerian public legal information. It concludes that the starting point towards the provision of adequate public access to the primary sources of Nigerian law is the abrogation of copyright in the texts of all government works, which include legal information, by amending section 4 of the Nigerian Copyright Act 1988 to reverse its copyright impediment to free and adequate access. This will introduce the necessary paradigm shift towards the realisation that governments hold public information in trust for the people who are its rightful owners and therefore have the right of free and adequate access to it. In addition, the enactment of appropriate legislation that clearly defines the duty of the federal, state, and local governments to develop and maintain comprehensive and up-to-date legal information online databases with free access is indispensable.


Akoma Ntoso for Accessibility of FAO and UN Resolutions

Abstract

Akoma Ntoso is an international Legal XML standard that has been approved by the OASIS body as a Committee Specification. It was originally developed to model legislative, parliamentary and judiciary documents using Semantic Web design principles. However, even other types of normative and regulative documents can derive the benefits of being represented in Akoma Ntoso (now AKN), since this language is able to describe their structure, their components (e.g., attachments or documents included inside of the main document), the references to and from other documents, the semantic annotation of some peculiar parts of the regulative language (e.g., actions, purposes), the workflow of the creation process, and the modifications over time in a formal manner. Thanks to such expressivity, FAO has modeled in AKN basic texts (e.g., Constitution), council reports, codex standards and resolutions since its first Session in 1945. From this deep theoretical and empirical analysis, conducted by the FAO team and CIRSFID, we have defined specific guidelines on the use of Akoma Ntoso in this specific scenario. Based on this work, UN High Level Committee on Management (HLCM) of the Chief Executive Board for Coordination (CEB) lunched a similar initiative in June 2016. They adopted a profile of Akoma Ntoso (AKN4UN) as the UN common standard in March 2017. We intend to show how Akoma Ntoso itself improves the accessibility of the FAO and UN resolutions using customized AKN schema.

Resolution Markup

The FAO and UN resolutions are regulative documents inside of each respective organization and some of them also have value for interconnected agencies. Usually, a resolution is drafted within a separate process and the official deliberation is nested, voted and published thanks to a Report document. For this reason, the legal analysis of the resolution determined that the <statement> document type of Akoma Ntoso was the best option.

Resolver of IRI

As every other UN agency, also the UN and the FAO publish documents in six official languages (English, France, Spanish, Chinese, Russian, Arabic). For this reason, we use Akoma Ntoso IRI naming convention to properly navigate the correct document in time and in the corresponding language.

Metadata and UNDO Ontology

The resolution includes several important metadata to qualify the portions of the text, among which temporal metadata. Purposes, Actions, Organizations, Persons, Roles are the main metadata to annotate the regulative purposes and the linguistic actions, the organizations and the persons involved in the resolutions (e.g., European Economic Community) and the related roles (e.g., Secretary-General).

UNDO (UN document ontology) integrates the metadata level with an ontology with the aim of managing changes over time (e.g., the change of nomenclature of the committees). Therefore, it is possible to query “give me all the resolutions between 2011-2018 with an action ‘noting’ provided by the UN acted by Secretary-General, with modifications”.

Conclusion

Akoma Ntoso is a completed and flexible legal XML vocabulary that is capable of annotating not only legislative documents, but also UN and FAO resolutions. The combination of Akoma Ntoso and UNDO ontology permits to enhance the searching capacity and the presentation accessibility. The AKN resolver permits to navigate the cross-references among UN document agencies and to access different languages.


Analyzing NJDG Data to Improve Access to Justice in Indian Courts

Abstract

Indian Judiciary is suffering from burden of millions of pending cases in its courts at all the levels. Hon'ble Supreme Court of India has initiated e-Courts project to deploy Information and Communication Technology (ICT) in the judiciary so as to efficiently impart justice without compromising on its quality. The National Judicial Data Grid (NJDG) is an important outcome of this project that indexes all the cases pending in the courts and publishes the data publicly. The launch of NJDG has also resulted in a jump of 30 ranks in the World Bank's Ease Of Doing Business Report. In this paper, we collected data on 44 days chosen randomly between August 31, 2017 to June 1, 2018, thus spanning more than nine months. Our dataset contains the statistics of the national level as well as of the lower courts in 32 states and 24 high courts. Our main finding is that the data present on NJDG is still very far from being in a state where it can be actually used by the Indian Judiciary reliably. There is a lot of scope for improvement in the statistics shown on the portal. This paper aims to improve the data present on NJDG and since data is collected over a period of nine months, we can clearly differentiate between temporally local and long term errors.

We also observe from this data that the scheduling of court cases should be improved to make better use of the available resources. We motivate the problem in detail and emphasize the need of designing efficient algorithms for scheduling cases in Indian courts so that the number of pending cases in courts reduce at a satisfactory rate. In the long run, use of machine learning and natural language processing will be required to improve upon the scenario.


Automation, Legislative Production and Modernization of the Legislative Machine: The New Frontiers of Artificial Intelligence Applied to Law and e-Democracy

Abstract

Electronic democracy is still far from being realized and several issues must be solved in order to make it possible. The quantitative problem of popular participation is one of them, but it can be mitigated through automation.

In fact, automated acquisition and processing of information is an essential component for e-democracy: it can help overcoming not only the just mentioned issue, but it can also assist lawmakers in carrying out their task.

Since the Seventies, the creation, diffusion, and technological development of computers and databases have allowed increasingly precise and refined processing of personal data exceeding the limitations inherent to those carried out with non-automatic tools. Similarly, the development and the use of advanced systems for processing particular types of information may enable the démos to be informed and effectively participate in the political life of the State, on the one hand, and lawmakers to fulfill their institutional tasks more efficiently, on the other hand. In other words, the legislative machine, regardless of its composition, can benefit from the automation of certain activities through the development and use of particularly advanced and “smart” IT tools.

This paper proposes two main applications that may help building a multilevel digital agorà where démos, lawmakers, governments, and public administration may cooperate.

The first is related to the integration, in each platform used for this purpose, of specific decision support systems, made up of a database of legal texts (laws and judgements) and of an expert system that allows people to understand the information stored in the database. The aforementioned systems should be usable by citizens (or legitimate users) in order to enable them to fully and consciously participate in a digital agorà thanks to the possibility of forming their opinion on a particular subject or case with the intermediation of a subject whose automated impartiality is guaranteed and verifiable.

The second is inherent in the use of IT tools that, integrated into a digital agorà, allow to transform the multiplicity of individual contributions into a general will; this may happen through a gradual reduction of the number of such interactions thanks to automated means. Moreover, using the same technologies used for the first application, it is possible to simulate, in advance, the effects of one or more decisions (just think of the participatory budget, in which each decision can influence the others).


Digitizing Luxembourg’s Legal Corpora: Experience and Vision

Abstract

An important thrust of e-Government is increasing citizens’ access to legal texts such as laws and regulations through online portals. While formats such as HTML and PDF have traditionally been the norm for these portals, there has been a rapid shift in recent years toward markup representations that provide legal information alongside the texts.

With the release in 2017 of the digital legal portal legilux.lu (http://www. legilux.public.lu), Luxembourg went a further step to provide an official and in-force electronic version of the texts, the paper version not having legal value anymore.

During the digitization effort, the need was raised to convert large amounts of consolidated “Codes”, available online only as PDF documents. An immediate challenge in this respect was, given the sheer scale of the legal corpora that need to be enhanced with metadata, a fully manual process would be extremely laborious and prohibitively expensive. This called for techniques that can provide automated assistance in identifying and annotating the metadata items.

In the past two years, SCL – the government agency in charge of legilux.lu – and the SnT Centre at the University of Luxembourg have been engaged in a collaborative project aimed at automating the extraction of metadata from legal texts. The metadata items currently covered concern the hierarchical organization of legal texts and their cross references.

As part of the above joint initiative, an automated assistance tool, named ARMLET (Automated Retrieval of Metadata from Legal Texts) has been developed in order to automate the conversion of legal acts. In particular, ARMLET leverages artificial intelligence technologies, particularly Natural Language Processing (NLP), for automated demarcation of legal metadata and converting legal texts into a structured XML documents. It has been successfully applied for converting seven of Luxembourg’s legislative codes into digital resources, most of which have been published on legilux.lu or are under review.

In our presentation, we will be describing ARMLET alongside our preliminary experience applying it to legal texts in Luxembourg,


Indigo Legislation Platform - Capture, Consolidate and Publish Legislation in the Cloud

Abstract

In this talk we present the Indigo Platform and the open source technology behind it, the challenges encountered developing it, the value that it unlocks for legal experts and the general public.

The Indigo Platform (https://openup.org.za/indigo) is an open source, cloud-based editor that simplifies the process of capturing, consolidating and publishing legislation so that it’s easier to read, work with and understand.

We are using the Indigo Platform to help promote access to the law at the national levels in a number of African countries, and at the municipal level in South Africa. AfricanLII uses the Indigo Platform to enable access to legislation in a growing number countries in Africa, including South Africa, Nigeria, Swaziland and the Seychelles. OpenUp is using the Indigo platform to help South African municipalities provide up-to-date, easy to read and share by-laws to residents, municipal staff and law enforcement.

Indigo can be readily adapted to support local legal traditions. For example, OpenUp recently assisted the ePaństwo Foundation (http://epf.org.pl) with adding support for the Polish legal tradition to Indigo.

The Indigo Platform captures legislation using the open Akoma Ntoso (AKN) XML standard. The legislation is stored in a machine-friendly fashion, allowing Indigo to understand the structure of the legislation. This means that Indigo can instantly format and publish legislation for desktop websites, mobile websites, e-readers (ePUB) and as PDF documents. This machine-friendly format enables new experiences that are not possible when legislation is treated simply as a Word document.

Indigo’s development is highly iterative and follows an agile software development methodology. We conducted extensive consultation with experienced legal editors to both understand the legislation consolidation process, and to identify opportunities to build modern software tooling to improve the process and to find new ways of solving problems. We regularly test Indigo with users that represent a range of editorial and technical skills. Updates and improvements are released daily in response to user feedback.

While Akoma Ntoso underpins the power and flexibility of Indigo, it has also presented challenges. Akoma Ntoso is more rigidly structured than other document standards most users are familiar with, such as HTML and Microsoft Word. We had to make trade-offs to strike a balance between supporting Akoma Ntoso’s rich structural representation, and providing a user-friendly editing interface that is readily customisable to local traditions.

The platform has a user interface that allows for easy interaction with the data, even for legal professionals. Its primary focus, however, is as a back-end application that provides simple APIs that allow Indigo data to be consumed by other applications.

The Indigo Platform embodies a simple notion - machine-readable, metadata-rich and versioned legislation is the accepted standard and best practice in the industry. The core focus of the Platform is productivity - automating data capture from a variety of formats, support for various editorial workflows, functions and production processes, while catering to diverse legal traditions.


EU Judicial Procedures and Case Law Databases: What’s Going On and What May Lay Ahead

Abstract

The raise of computational power, the boost of electronic data storage capabilities and the growing ubiquitousness of the Internet, have facilitated the collection of legal information, and increased its availability for stakeholders. At European level, access to European and national legislation and case law is seen as being of paramount importance for economic activities and for upholding the rule of law. EU institutions and key stakeholders are seeking to support initiatives that go in this direction. After creating dedicated portals and databases to facilitate access to EU legislation and case law (e.g. EUR-Lex, Curia), the goal to support the establishment and concrete use of different European legal instruments that aim to facilitate access to justice in cross-border situations make further steps necessary. This paper investigates existing electronic databases that have been created to disseminate case law information related to the application of EU judicial procedures. In doing so, it explores the potential these databases create for an improved application of European procedural instruments, forwarding their use and creation of a common understanding.

The paper begins also to consider an important novelty that can represent a new opportunity in developing more integrated cross-national legal databases by making use of the increased availability of electronic procedures at national, and, lately, also at EU level. This technology and law development may open up and facilitate the collection and availability of national court decisions. These developments result in digital by default judgments, which are the outcome of electronic procedures. At EU level, in the e-CODEX pilots, the EOP and the ESCP court decisions can be generated from electronic filed standard forms and structured data exchanged during the procedure. This implies that, in a database alimented by electronic decisions which are not just text but include structured data generated during the procedure, many of the element of the case could be used to support ‘smarter’ research for practitioners and interested parties. This may significantly reduce the need of subsequent work related to the classification of data by experts in the field or the task of anonymizing specific fields. Additional data generated during the procedure, but not available in the sentence could also be used to enrich the database. Furthermore, as the support of the procedure is multilingual and e-CODEX supports semantic interoperability, all the structured data would be multilingual by default.


Emancipating Law and Legal Knowledge by Government and Non-State Actors in India

Abstract

The contribution is an initiative to envisage efforts of Government of India and Non-Government Organizations envisioning legal information available in public domain and generated by public authorities. India is the 3rd largest Economy of Asia and setting new cornerstones in the field of law and legal research.  From ancient kings’ rules to the largest democratic system of the world, Indian law has been developed to serve more than 1.27 billion of its citizens. India have more than 23 Law schools/universities and 1500 law colleges/departments.

Right to Information Act, 2005 have empowered citizens to get public related information explicitly. However, National Data Sharing and Accessibility Policy (NDSAP) was drafted and introduced in India in 2012.  For making available information for public for social, economic and business development. Policy supports visualization of public information through www.data.gov.in. A detailed guideline for National Data Sharing and Accessibility Policy has been prepared by the Open government Data Division, National Informatics Centre, Department of Electronics and Information Technology, Ministry of Communications and Information Technology, Government of India in November 2015.  The resources supported is well known as “Open government data (OGD) Platform India” (OGDPI) with major objective to publish/visualize data sets, reports, service procedures, dissemination and collection of data by the identities of Government departments/ ministries/autonomous organizations of India.  

Apart from it, a number of other governmental web portals viz. The website of Parliament of India (www.parliamentofindia.nic.in), INDIA CODE (www.indiacode.nic.in), Judgement Information System (JUDIS) (www.judis.nic.in), COURTNIC (www.courtnic.nic.in), INDIAN COURTS (www.indiancourts.nic.in), National Human Rights Commission (www.nhrc.nic.in) and Law Commission of India (www.lawcommissionofindia.nic.in), are supported by National Informatics Centre for facilitating free access of legal information.

Legal Information Institute of India (www.liiofindia.org), Indian kanoon (www.indiankanoon.org) Legal Service India (http://www.legalserviceindia.com/) Legal India (http://www.legalindia.com/), Parliament Research Studies India (www.prsindia.org), and Commonwealth Human Rights Initiative (http://www.humanrightsinitiative.org/) are other valuable initiatives which facilitates its precious supports for accessing and disseminating public legal information pertaining to India. The paper buttresses of sound information system conjured by various initiatives supported by information technology for development of efficient legal system succoured better justice.


Free Access to United Nations Documents and Publications: Strategies for Navigating a Complex Landscape

Abstract

Even if one could think that all the documents and publications edited by United Nations are at disposal for free on the web, unfortunately this is not true. Digital-born (issued since 1993) or digitized official documents, having a United Nations Symbol, are at disposal at on-line institutional data-bases like O.D.S. (Official Documents System), or its upgraded version the: United Nations Digital Library, as also public domain publications. But for sales publications is different, they are not always at disposal on electronic format, nor at disposal for free. Digital copies may be available on institutional web-sites, in most cases only recent issues are present, for free, on web-pages dedicated to publications. Those dedicated pages have to be considered separately from the above mentioned on-line data-bases, not only because it’s rare that the contents corresponds (in principle data-bases are for public-domain publications and institutional web-pages for sales publications), but also because on-line data-bases’ aim is not to put at disposal digital versions of sales publications, but to enable searching contents inside publications, or simply find out the existence of a publication, the United Nations Digital Library is first of all a bibliographic data-base, giving access also to some files. The difference is subtle, but meaningful for publications, because sales publications are serials, or publication with a periodical nature, their issues are published once to four times a year, and besides their existence and contents, also their “periodicity” has a meaning, in fact you can “spoil” a periodical also reading through its evolutions during time. For this reason also paper copies not yet digitized keep their value, as the evolution of the serial title during time, and the reconstruction of this evolution. Since two years ago, to this complex network of documents and publications on-line the United Nations Board of Publication added another instrument: the institutional Repository called Un-i-Library, which is defined as “a comprehensive source of digital contents”, inside you can find sales publications, as various yearbooks, for instance the United  Nations Yearbook, and other serials as the “Treaty series”. The repository is accessible for free in the e-read format, against subscription for the e-pub and pdf versions.

The United Nations documents and publications digital landscape is so complex that you need a guide to navigate it, this is the role of the two main United Nations libraries in New York and in Geneva, but also of the United Nations Depository Libraries present in different member states, whose aim, along with keeping paper copies of documents and publications is to help users to have access to United Nations documents and publications. Although, in order to afford the on-line challenge, those libraries have to change the way they participate to  the United Nations Depository Libraries System, both offering new services than keeping the heritage they collected during years of participation to the system and to do this they need an effective dialog and cooperation with the United Nations Board of Publication that issued the on-line institutional repository and the two United Nations Libraries.


How to Create an International Encyclopaedia of Case Law

Abstract

How to create an international encyclopaedia of case law

Would it be possible to create an international encyclopaedia of case law?Legal practitioners are often required to consult case law from other jurisdictions. The development of the internet means that it should now be possible to create an international publication with short reports on all written judgments - an international encyclopaedia of case law. Since 2011, this author has developed an online publication entitled Stare Decisis Hibernia, which has summarised and categorised over 7,000 written judgments of the Irish superior courts. A model and template have been developed that could be scaled up to an international level - to include both national courts and supra-national courts such as the Court of Justice of the European Union.

There would be challenges involved in making such an encyclopaedia international, in that some types of case are more common in some court systems than others, and some cases are significantly more technical than others. However, most cases can be categorised and explained with reference to the rights that the parties are seeking to vindicate.

The cost of preparing an entry for every written judgment would not be unduly burdensome, compared with the cost of the court system. If the encyclopaedia were to be available free of charge, it would need funding from the government, the court system or the legal profession. If such funding were not available, it could be funded by subscription - which is the model adopted by Stare Decisis Hibernia - at quite an affordable level for legal practitioners and students alike.

Further challenges would arise if the encyclopaedia were to be available in more than one language. It would be extremely expensive for example to translate each entry into every language of the European Union. However, selective translation would be a practical option with benefits to many users, and it could be funded by relevant international organisations.

In summary, an international encyclopaedia of case law would be of benefit to the legal profession worldwide. It could be prepared at a reasonable cost, following a model similar to that of Stare Decisis Hibernia. It could be funded by states, the legal profession or by subscription. It could be made available in more than one language. The only obstacle is the will to do it.


Legislation Supporting System Based on Legal Computer Programming As SDL (Software Defined Law)

Abstract

This research is to develop a system for supporting legislation. Most ordinances in local governments are not drafted by experts of law. In a lot of cases, such a drafter who has never learned law writes ordinances often. Moreover, the technique of drafting laws is like craftsmen's traditional one. At least in Japan, there are so many rules for drafting law that the manual of more than 500 pages cannot also cover them completely. These rules cannot be learned easily. Therefore, IT support would be very helpful for law-makers.

To implement the supports, the authors have studied and developed as an e-Legislation project for 10 years. Since 2009, showing the similarity between the process of software development and the process of legislation, the authors have advocated applying the methods of software development and their development tools to legislation. In 2012, based on the theory, the authors released the “eLen” ordinance database system to support legislation for all of Japanese local governments. This system has the functions for clustering similar ordinances and making comparison tables of ordinances. It has used in more than 770 governments so far. Since 2016, the authors have also studied the research for drafting laws described in computer programming languages. This report is a part of these researches and the most important core of them.

In this study, we propose the new system called “Xelen.” This system makes it possible to simulate laws and to draw the structure of laws systematically. Furthermore, it can bring us to draft correct style articles of laws from a program. Of course, law drafters as end-users do not write directly such a program but our editor system with the web GUI helps the users to write the program. Perhaps, they will use the system without recognizing that they are writing any computer program. They only input some words or items into the empty boxes in some templates of laws. Recently the authors also start to develop a system for auto-composing these templates by AI.

Different from the conventional “AI and Law” typically representing laws in logical formula, we adopt Python, which is a popular programming language. Based on input specification for these python programs of laws, the authors designed the following two subsystems of parts of integrated environment system for legislation.

  1. Legislation Compiler
    This is a subsystem to compile python programs of a law to the sentences of articles included in the law.
  2. Legislation Editor
    This is a subsystem to provide GUI for end-users drafting laws with law templates and icons as building blocks of laws.

By using such a system, the authors recommend systematic legislation process based on computer programming of laws. This approach can be called SDL (Software Defined Law) since operation of programs as laws makes it possible to simulate and verify these laws even if these cannot execute in our real world before enforcing laws.


Public Access to Primary Legal Information in Ghana: Opportunities and Challenges

Abstract

Information to a large extent can have economic, political and social impact on the lives of people globally. Information enables one to adjust in the global system through the effective utilization of local resources. People’s right to access information is mostly interpreted in the legal frameworks of countries that supports freedom of information. The citizenry needs to have access to information that is produced by governmental agencies.

It is expected that the availability of information in any democratic state, should allow the citizenry to freely express themselves, equally participate in the democratic governance of their states and have established equal opportunities for airing their views, for a more transparent, accountable and an open society.

The objective of this article was to establish and assess Ghana’s legislation on making legal information readily accessible to the general public especially at the national level. Access to public primary legal information is important. In Ghana, certain features in the Constitution of the Republic of Ghana have stipulated the human rights that the citizenry can enjoy and this is clearly stated under the Chapter 5 of the Constitution. Provision has also been made to the Right to Information as well as the academic freedom elements in Ghana’s 1992 Constitution under article 21(1)(f). One major attempt by the government to ensure that government data is made readily available to the public re-use was through the Ghana Open Data Initiative (GODI).

Other channels that enable the citizenry to have access to legal information includes platforms such as E-Governance, Research4Life Partnership Program through GOALI and the advent of Open Source Software. However, access to primary legal information in Ghana is beset with some challenges which includes intellectual property, funding, control of cyber space and the lack of needed local personnel with requisite skills. Some recommendations have been made based on the challenges identified.


Artificial Intelligence at Lexum

Abstract

Over the course of the last year Lexum has started exploring the potential of deep learning (DL) and machine learning (ML) technologies for legal research. Although these projects are still under the umbrella of Lexum’s research and development team, or Lexum’s Lab (https://lexum.com/en/ailab/), concrete applications have recently started to become available.  This paper explores two of these projects: Lexum’s Citation Predictor, and Lexum’s Learning to Rank solution.

These projects benefit from a combination of three factors. First, the millions of legal documents available in the CanLII database in parsable format along with structured metadata constitute a significant dataset to train AI algorithms. Second, Lexum has direct access to the knowledge and experience of one of the leading teams in AI and deep learning worldwide at the Montreal Institute for Learning Algorithms (MILA) of the University of Montreal.  Third, the availability of computer engineers with cutting-edge expertise in the specifics of legal documents facilitates the transition from theory to practical applications.

This papers first look at why the data available on the CanLII website is particularly well suited to train ML algorithms.  More specifically, the fact that its citator recognizes over 16 million citations has contributed to create a “map” of the Canadian Law. 

Regarding concrete outcomes, Lexum’s Citation Predictor can predict the most relevant sources of law for any given piece of text (incorporating legal citations or not).   Lexum’s approach consist in learning from the citation network to predict which sources of law are relevant to the text of a legal brief, a legal opinion or to the plain language description of a legal issue.

The Learning to Rank algorithm for its part improve the relevance of search results by learning from document structures, user behaviors and the content of the citation network. It weights-in all of these signals using a ML algorithm to better rank search results.


Building Common Datasets for Legal AI

Abstract

Open Law “AI and law – learning datasets” program aims to co-create learning datasets for AI algorithms applied to law: the scenario for this experimentation deals with the semantic structure of a court ruling. We state it is a profitable and essential basis for many other business-oriented cases, and enables a more reliable enrichment of case law contents. Moreover, the process itself and the methodology in the making came to be as valuable as the actual outputs, providing jurists involved in the project with a state-of-the-art machine learning experience.  

Machine learning revolutionizes the way artificial intelligence tools are designed, opposite to rule-based systems used in the legal domain before. As INRIA states in its white paper on AI, "the most remarkable results achieved in the field of machine learning are based on supervised learning, i.e. learning from examples in which the expected result is provided with the input data. This involves labelling the data with corresponding expected results, a process that requires large-scale data." In the legal world, particularly in France, these learning data do not exist, which entails an enormous entry cost for all players wishing to enter this emerging market. Moreover, this lack of quality training data is at the origin of many of the ethical risks (lack of transparency, amplification, reproduction or even creation of bias...) linked to the development of artificial intelligence tools.

This is the reason why the association Open Law launched a program “AI and law – learning datasets” –, in partnership with the CNIL's "Ethics and Algorithm" mission, which proposes to use the various legal and technical skills of the Open Law* association community to constitute several sets of training data, which will be usable by all in an open format.

In accordance to the "do-tank" objective that the association set for itself, the 2017 season of the AI and Law - Learning dataset program aimed to constitute in a restricted time length (6 months) a first set of data useful to as many people as possible (the zoning of court decisions, see below), and in doing so to:

  • demonstrate the feasibility and utility of building quality training data sets ;
  • explore methodologies for creating learning data and the tools and skills needed;
  • document the process to make it replicable for other data sets.

The sponsors of this 2017 program were: Open Law, CNIL's "Ethics and Algorithm" mission, ELS and the French judiciary Supreme Court (Cour de cassation). Other entities involved were: the French Ministry of Justice, SVP, IMT Atlantique, M2 Droit technologies du numérique et société de l’information (University Paris OuWest) and Natixis.


Entropy in Digital Information and the Enforcement of Law: Towards a Unification of Remedies?

Abstract

The term “entropy” measures the amount of energy that is unavailable within a closed system, with the aim of explaining the tendency to reach a state of thermal equilibrium.  Similarly, in information theory entropy is generally defined as the average amount of information produced by a hypothetical source of data.

Entropy refers in general to a state of disorder or uncertainty - hence the greater the entropy, the less information is sent or received, which in turns reduces the clarity and the unambiguous nature of the message.

This research aims to examine the legal remedies - both judicial and non-judicial - available in the area of electronic communication, adopting as the main comparison parameter the problem of the legal status of digital information.

From whichever conventional research perspective the problem of the enforcement of law on the Internet is framed (i.e. the protection of fundamental rights, of economic or personal interests, of particular categories of subjects or legal assets, the contractual or non-contractual relationships, the type of effect pursued by the chosen remedy and its true effectiveness etc.), the interpreter is often forced to search for and attribute meaning to a given event that hardly ever leads to clear-cut answers, either synchronously (with reference to the same occurrence) or diachronically (in the sense of analogical application of a given rule to a similar event, or one that is in some way classifiable within the same type).

The infocentric structure of today's society on the one hand does not allow for the advance identification of a clear and generalized correspondence between a subjective legal situation and digital information; on the other hand, protection mechanisms tend to converge both from a classification and a technical profile. In other words, the consolidated subjective right vs. remedy model - understood as a system of subjective situations that are pre-established by the law from which owners derive their faculty or powers and which puts the the obligation to do (or also not do) in the hands of individuals or the rest of the community, and alongside which a range of protection instruments can be found that can be invoked before the courts in the case of violations (ubi jus, ibi remedium), is often diminished and becomes more typically an action-reaction model.

It seems then useful to conduct the examination by placing the emphasis on the type of remedy available, evaluating firstly the 'resistance' of the classic two-way split into protections that directly and materially affect the circulation of digital information (comparable in the broad sense to the so-called property rules) and rules that oversee the ex post allocation of obligations of a compensatory nature (liability rules).

In a multi-subject context marked by a post-industrial, cognitive economic model, it is possible that at the operational level the administration of one type of remedy implies a different consequence for all the other subjects involved in the information flow. While respecting the diversity of the experiences analysed, the regulatory trend seems to be that of the parcelling up of behavioral standards in a preventive and collaborative key.


International Environmental Agreements in the Age of Big Data and Algorithms

Abstract

Big data is considered as the new oil due to the fact that such data is used for making a great amount of money by tech companies. However, with the problems that are be ingnoticed in the private sphere, there are new questions emerging and revolving around the potential effect of big data and algorithms in the environmental field from a legal perspective. In fact, since big data and algorithms have become more and more relevant in the environmental field, one would wonder whether international environmental agreements should start taking this fact into consideration through the incorporation of new provisions in these agreements tackling these matters. To answer this question, this article will first discuss the concept of big data and algorithms as applied in the private sphere. Then, a brief overview of the ways in which big data is being introduced in the environmental field will take place. Later, the issue of whether international environmental agreements should incorporate these matters in their provisions shall be examined where the current duty of information exchange will be studied and a specific emphasis on transboundary water agreements will be made.

The final section of the article will attempt to provide a suggestion for the incorporation of new provisions in international environmental agreements that adequately tackle big data and algorithms. To do so, the focus shall be on several provisions of the regulation 2016/679 of the European Union (EU) entitled General Data Protection Regulation (GDPR).


Legal Limitations of Algorithmic Analytical Tools Against Disinformation

Abstract

The spread of fake news, disinformation and anti-democracy propaganda through the online media and social networks is an acute problem. Particularly challenging is the need for credible, reliable and scalable approach to swift identification and assessment of illegal content that is necessary for legal process of takedown or removal by the responsible internet service provider and also for subsequent criminal prosecution of the offender. The author of the contribution is a member of an interdisciplinary research team of IT experts, security analysts and lawyers focusing on developing a method of recognition, detection and analysis of manipulative techniques on pro-Russian online media outlets through automated analytical software based on machine learning and big data mining [Project MUNI/G/0872/2016 Manipulative techniques of propaganda in the age of internet].

The use of an analytical tool for recognition of propaganda disinformation, as the one that is being developed in the aforementioned project, evokes several important legal questions that need to be considered in order to properly assess the potential viability of this approach in the struggle against propaganda manipulation. The particular issue at hand that is to be discussed in this contribution are the parameters of the legal framework for this analytical software and the ensuing legal limitations of this tool against disinformation. This form of data analytics should provide evidence for criminal or civil liability procedures, but such application requires critical probes into objectivity and impartiality of the algorithmic tool. To pursue this issue to its core, it is suitable to follow to the basis of Wiener´s cybernetics and identify the limits of compatibility between application of law and algorithmic expression of abstract terms. Patterns, correlations and axioms need to be identified and verified in order to imitate legal or other abstract expert assessment. Such inputs into the algorithmic decision making need, however, to be subject to adequate scrutiny and verification, in order to prevent from flawed or biased foundation for interpretation and application of law. This challenge can also be found in the background of many legal frameworks aimed at establishment of tools necessary for appropriate regulation and moderation of online conduct and data flows.

The contribution shall discuss the presented challenges in greater detail as well as provide the summary of findings obtained on the current pursuit towards mitigating the limitations emerging in their consequence. The aim is to provide a framework of an appropriate setting of the analytical algorithm that would satisfy the legal prerequisites for producing supporting evidence in a process of takedown or removal by the responsible internet service provider or in a criminal prosecution of the offender.


Opportunities and Challenges in the Legal Tech Services in the Italian and European Framework

Abstract

It is undeniable that legal tech services can render a country’s legal system more efficient, providing greater transparency and rapidity in the decision-making process. These services are however seen by traditional law firms as competitors. Nevertheless, legal tech services should not be viewed as competing with law firms, but instead should be conceived of as the future of the legal profession. It is beyond doubt that the digital revolution has changed many market sectors (consider for example the tourism or transport sectors). Despite this, the legal profession is still closely tied to a fiduciary relationship which cannot be affected by artificial intelligence, provided that there is willingness to accept that certain legal activities which are not particularly complex (for example a lease agreement for a building) must inevitably be automated. Once this type of approach has been accepted, the legal profession of the future will be essentially based on a lawyer’s strategic ability to choose the best from amongst the various solutions put forward by artificial intelligence.

This aspect can be summarised as the “Mida Paradox” according to which “Machines take the law literally...  Humans don’t”. Within such a wide, complex panorama of information, it is quite possible that there are legal content losses in algorithmic translations of regulations. Any interpretation of the law presupposes not only knowledge of implicit information which is not contained within a document to be analysed by artificial intelligence, but also the ability to perform assessments on the basis of a level of logical abstraction which it is difficult for a machine to attain. For example, any decision regarding a criminal defendant’s mens rea on the basis of whether or not he wishes to commit particular type of offence entails an extremely complex decision-making process which is very hard to translate into an algorithm.


Using Machine Learning to Predict Success Or Failure in Chapter 13 Bankruptcy Cases

Abstract

Obtaining a chapter 13 bankruptcy discharge in the United States is notoriously difficult. Past empirical studies conclude that only one-third of chapter 13 debtors complete their obligations under their plans and obtain a chapter 13 discharge. About 60% of cases end up dismissed, or converted to a Chapter 7 liquidating case.

In this project the author examined a public case level database made available in 2017 by the US Federal Judicial Center, based on information collected by the Administrative Office of the United States Courts. The author obtained from the Federal Judicial Center’s Integrated Database the complete database of cases pending between fiscal year 2008 through 2017.  Based on preliminary analytical work, the author generated a reduced set of about 743,000 Chapter 13 cases that were originally filed during fiscal years 2008 and 2009.

These records, representing substantially all of the Chapter 13 cases filed during those two years, were randomized, and smaller training and test sets extracted. From these sets, “features” were engineered suitable for use by machine learning algorithms. A success feature was created, based on whether a debtor obtained a discharge in bankruptcy and was still in a Chapter 13 bankruptcy case at the time of case disposition. Against this label, the model examined:

•     whether the debtor had an attorney;

•     whether the debtor had filed a prior bankruptcy case;

•     whether the debtor filed with a spouse or not;

•     whether and how the debtor paid the case filing fees;

•     the gap between the debtor’s income and expenses;

•     the ratio between the debtor’s income and expenses;

•     whether the debtor owned a house;

•     whether the debtor provided information about the amount of his or her unsecured debts;

•     the debtor’s income, debt amounts, and the value of both real property and personal property owned, all as reported in the initial filings for the case; and

•     the judicial district in which the debtor filed his or her case.

Three algorithms were tested: a logistic regression classifier, a neural network, and a random forest decision tree. The best results were obtained using the random forest decision tree built on a training set of  420,000 bankruptcy cases. Run against three test sets of 30,000 cases each, the model achieved accuracies of .6978, .6991, and .7033. AUC’s were all slightly above .70. These predictions exceeded a baseline of about 62%. The model also shows very high predictive ability for about 25% of cases, when using generated probabilities of success or failure. By extrapolating the model’s probability outputs, the author was able to precisely predict the percentage chance of success for individual bankruptcy cases.

A detailed paper is accepted for publication in the 2018 Norton Annual Survey of Bankruptcy law and will be available on Westlaw in August 2018.