-

Big Data

Legal Analytics - Law & data science is much more than privacy alone

DSC/t Blog - March 2016 | Corien Prins - Dean Tilburg Law School, Professor of Law and Informatization

“Once we have fully artificial intelligence enhanced programs like LegalZoom, there will be no need for lawyers, aside from the highly specialized and expensive large-law-firm variety.”


A thought-provoking quote from the August 2014 issue of ABA Journal, the flagship magazine of the American Bar Association.

Will in the not-too-distant future intelligent data-based litigation assistants indeed have the ability to reduce legal advice to the simplicity of performing a search? Imagine a responsive engine, assembled from Google’s proprietary data collected from online searches, that informs people on a case at hand (background details on relevant parties, the facts, the merits, background on courts and judges, the remedy sought and share any relevant documents). Based on an advanced algorithm that mapped out the relationship between all of the relevant case law, the application could generate forecasts of how the case would be resolved with different courts and even different judges, and perhaps even recommend an ideal venue. The application could explain how best to structure the litigation, what types of motions would be most successful, and how to arrange arguments. With advances in artificial intelligence, it is not difficult to conceive of an intelligent system even drafting the briefs.

Legal Analytics
Legal research in the domain of big data is primarily focused on what can be described as ‘the legal aspects and implications of …’, i.e. issues related to privacy, security and property of data. Far less attention is given to the role and impact of predictive tools on legal decision-making itself. Sufficiently large datasets, the right analytical tools and machine learning techniques potentially promise answers to traditionally difficult questions such as: where are legislative resources most effectively employed? How do patterns identified by means of analytics guide decisions of judges? In what domains can tools be developed and applied to predict the effectiveness and legitimacy of regulatory interventions? What datasets and mining techniques are most suitable to discover trends and patterns on the interaction of large audiences (groups of citizens) with regulatory measures? Would the use of visualization enhance the engagement of citizens with regulatory policymaking?

In the United States developments have taken a rapid pace, with an important role for private companies. To illustrate: Lex Machina is a private analytics company aiming to predict the cost and outcome of intellectual property litigation, a predictive model has been developed to track settlement outcomes of securities fraud lawsuits and another application is used in electronic discovery to decrease the costs of civil litigation.

As far as the Netherlands is concerned legal big data awareness slowly comes to the surface. For example, the Dutch Legal Tech Meetup, an initiative of Tilburg University alumnus Jelle van Veenen, has placed the potential of data science and artificial intelligence on the legal agenda. Another indication of the growing interest for big data is a special edition of a Wolters Kluwer newsletter for Dutch law firms on big data developments and innovation for legal practioners.

Four domains
In law and legal practice roughly four domains can be detected where there is a need for data science expertise as well as legal professionals that know about the potential as well as disruptive effects of data science.

Courts

Data scientists can offer expertise on how informed legal decisions can be made in using data about the relevant precedents, the issues in a particular case, and a judge’s profile. Illustrative for the potential innovation in this domain are research results published in 2014, where sophisticated techniques from machine learning were applied to predict the voting behavior of the U.S. Supreme Court. Using only information available prior to the court’s decisions, their model was able to correctly predict 70 percent of the court’s ultimate outcomes.

Data scientists can assist judges in crawling through submitted evidentiary material, categorize it, and build a profile on the relevant issues to be ruled on. Data scientists can also develop and deploy services that allow for intelligent analyses - using (publicly available) data – relevant in court procedures.

Interesting aspect coming to the fore in describing these data science tasks is whether or not one can guarantee the same level of transparency and accountability as is now asked from judges? For it is the requirement of transparency that – combined with accountability - ensures that court rulings are clear, comprehensible, and susceptible to objection. Also, what level of accuracy is required if we are indeed to value humanity over accuracy in the dispute resolution process? Problematic here is that today’s search and recommendation algorithms rarely explain their outcomes, thereby leaving it to users of the technology to guess what the underlying reasoning might be, which is a poor foundation for well-informed, accountable decision-making in a democratic society.

Legal practice/aid

Clearly, there is a need for data scientists having expertise on the use of pattern recognition in legal advice, complaints and electronic transcripts of court procedures. Experts knowing how to develop and use intelligent systems that are able to understand what a client’s legal issues are by connecting the legal rules—as determined by the algorithms—and the fact patterns offered by someone in need of legal services. Experts knowing how to develop legal referral websites, measuring value for money in legal services and review sites where citizens choose to review lawyers, judges as well as legal advisors. Experts knowing how to develop new services similar to those already offered by US-companies.

Last but least one can also think of data scientists assisting lawyers in due diligence procedures.

Legislature and policy makers

Our society also needs data scientists having learned how prediction (and visualization) may assist to address the complexity of legislation and regulatory initiatives. How the smart use of data may help predict societal acceptance of legislation. We need e.g. data scientists with expertise on how innovative systems may provide advice on what an administrative agency, or zoning board, etc. will do with a particular case. Or experts who know how to use data science for criminal investigations, tax fraud detection and the use of intelligent systems to assist government officials in public administration matters.

Private companies

In 2015 the start-up Clocktimizer won the Legal Tech Start-Up Award. On its website, the company states “it helps law firms to unlock the full value of their timetracking data including the narratives. Clocktimizer’s insights enable law firms to improve their pricing, to control budget and scope and to increase customer satisfaction with full transparency on fees.”

But there is more than helping out law firms with data science applications. Clearly, data science can also offer established as well as new companies a business case in combining big data and advanced algorithms that enable company systems to automatically aggregate and organize contractual clauses based on legal texts. And thus we need experts knowing how to deconstruct legal documents by analyzing data to facilitate company lawyers in contracting and judging risks and contractual complexity.

Minor on Legal Analytics
Data science is likely to act as a disruptive as well as an enabling technology in that it affects not only the production and delivery of legal services, but also the legal infrastructure, legal business structures as well as the legislative process.

Many challenging questions arise. E.g.: by analyzing large amounts of data, predictions can be made on how a judge, or particular court, will resolve a particular case. But what if a judge deviates from what is expected? Could the tool undermine the legitimacy of the court? On the other hand, when algorithms can predict what judges do numerous citizens who today cannot afford access to legal services will have the opportunity to pose questions and receive answers from an algorithm that is entirely automated, and may only need to bring an attorney into the picture under certain more difficult circumstances. Hence, what do we lose by the absence of human interference? Do we value humanity over accuracy in the dispute resolution process?

The minor “Interfacing data science and the law", part of the bachelor program offered as of September 2016 by the universities of Eindhoven and Tilburg will introduce data science students into the world of the judiciary, litigation, legislative power and what data science could mean for the legal profession.

Tilburg Law School is responsible for this minor and currently developing its program. Different law firms and actors in the legal domain have indicated their interest in participating in the courses and offering a learning environment for students. In addition, researchers from Tilburg Law School join forces in researching the many challenging questions that our society, legal professionals, policy makers, academia and many other actors face given that data science will definitely influence the legal domain and the legal and policy-making profession.