Throughout the twentieth and early twenty-first centuries, the world witnessed exponential growth in the amount of information collected, stored, and analyzed. Not unrelatedly, the cost of storing, analyzing, and transmitting information has fallen exponentially during the same period, resulting in substantially greater access to and awareness of data. A few hundred years ago, the cost of storing the sum of written human knowledge exceeded the budgets of all but the very wealthiest institutions; today, the body of all written human knowledge can be accessed by a device that nearly half of the world’s population carries in their pocket. As a result, people have become somewhat desensitized to the purpose and presence of data in modern society. In order to understand the purpose of the data and knowledge we have accrued and stored – and which is intrinsically linked to informatics – it is useful to look back to a time when knowledge was genuinely scarce and expensive to access. To understand informatics, we must understand the importance of writing and knowledge systems through history.

Understanding what happens at scale in our judicial systems seems a relatively simple problem, but has proved difficult in the past due to the inability of practitioners to access the needed information. In this chapter, we examine bankruptcy case information made available by the Federal Judicial Center in 2017 and use that data to understand the difficulty Chapter 13 bankruptcy filers have in obtaining their bankruptcy discharge, the potential factors that correlate with obtaining a discharge, and to predict how likely a specific case is to succeed. We discuss a project conducted using data from over 700,000 cases, as well as examine much smaller data sets suitable for manipulation using Excel.

The law is not free in either sense of the word: It is costly to access, and there are restrictions on its use. An ideal system would allow any member of the public free online access to all official primary law – cases, statutes, and regulations. In the twenty-first century, a digital-first court publishing system, and primary law in general, should be online, free, open, comprehensive, official, citable, and machine-readable. A modern online law system should also have eight additional, more specific characteristics: digital signing, versioning, good structure, medium-neutrality, archives, a search function, bulk conveyance, and an application program interface (API).1 Unfortunately, such an ideal system remains a distant aspiration in most jurisdictions.

Document assembly, the computer-assisted generation of texts, is a form of automation with wide applicability in law. Most legal tasks involve document preparation. Drafting effective texts is central to lawyering, judging, legislating, and regulating. The best way to support that work with intelligent tools is an ancient topic in legal informatics circles. This chapter surveys the history and current state of the legal document assembly field and reviews the typical concepts, features, and development processes involved. It describes illustrative applications of document drafting software and the knowledge representations and interfaces they leverage.

Somewhere between sonnets and Beat literature lies the notion of the semi-structured document, in which some type of, if not formalized, than perhaps anticipated, or at least understandable, textual layout, sectioning, and various forms of visual and labeled elements assist the reader in deciphering the document’s meaning. In fact, had the publisher been able to accommodate it, Faulkner had hoped to publish his masterpiece, The Sound and the Fury, in various colored text, each color denoting a different period of time, in order to assist the reader in making sense of what might otherwise be viewed as incoherent ramblings coming through his severely mentally disabled character’s narrative.2 And such is the current fate and ongoing dilemma facing a legal system that has historically been centered on the static, written document, collecting dust if lucky enough to be on a shelf, rather than buried in a random box under a desk in a remote office, yet all the while holding forth the rules and consequences of not abiding its every word and punctuation mark, scribbler’s errors included. This is the glue that holds society together.

The past decade has seen an increase in conversations about innovation in law, with a focus on how new technology can make the legal system more efficient and effective. Legal technology, in the form of artificial intelligence (AI), data analytics, and mobile applications, has been heralded as bringing a new era of legal services. This chapter advocates for a distinct but complementary approach to legal innovation, based in human-centered design, which can create new non-technological innovations and improve how lawyers and laypeople engage with legal technology. If a technology-driven approach focuses on how to make systems more intelligent and more efficient, a design-driven approach focuses on how to make systems that people want to use, are able to use, and that give them value.

The storage, description, collection, organization, and selection of legal information is central to a legal system, whether public or private, professional or lay person, structured or unstructured, textual or non-textual. The ability to get timely information where it is needed is as important to law as it is to any other field, given that law is a system of written rules and procedures. Ignorantia juris non excusat – ignorance of the law is no excuse. But where citizens seek to know the law and cannot find it, they have a right to question its legitimacy, at least as applied to them. The function of the reference librarian within a brick and mortar library, while still invaluable, does not scale to the magnitude of the problem. Technological information intermediation, such as a search engine, can thus be increasingly viewed as a necessary component of a modern legal system, and familiarity with the basic concepts is an important tool in the legal technologist’s toolkit.

The philosophy of law is the study of the nature of law: What is law? What are the criteria of a functioning legal system? What is the relationship between law and morality? A course on legal technology and legal informatics focuses on the technological implementation of a legal system. As we move away from static, printed documents toward virtual, distributed, integrated systems, software (“code”) plays an increasingly important role in the legal system. Obviously code has applications far beyond implementing law. Yet, as Lawrence Lessig points out in Code 2.0, non-law code also regulates behavior, often in a more fundamental way than laws do.1 As discussed in this chapter, code is architectural in nature and effectively limits behavior similarly to laws of physics. Only in science fiction do we entertain the notion of faster-than-light travel, perpetual motion, or evading gravitational forces. Similarly, we tend to accept the limitations of code that prevent us from, say, lending out our e-books, though such lending would certainly be legal. Even if we are aware of these limits and do not like them, most of us have no capacity to change them. As far as our behavior is concerned, code may as well be the law.2

Almost all law is expressed in natural language; therefore, natural language processing (NLP) is a key component of understanding and predicting law. Natural language processing converts unstructured text into a formal representation that computers can understand and analyze. This technology has already intersected with law, and is poised to experience rapid innovation and widespread adoption. There are three reasons for this: (1) the number of repositories of digitized machine-readable legal text data is growing; (2) advances in NLP tools are being driven by algorithmic and hardware improvements; and (3) there is great potential to dramatically improve the effectiveness of legal services due to inefficiencies in its current practice.

This case study describes how a team of computer scientists assisted a team of public health researchers by applying machine learning to extract information from statutory texts. Researchers at the University of Pittsburgh’s Graduate School of Public Health (SPH) had been manually mining specific information from federal, state, and local laws and regulations concerning public health system emergency preparedness and response. The analysts used the information to assess and compare states’ regulatory frameworks concerning emergency preparedness. They retrieved candidate legal and regulatory texts from a full-text legal information service, identified relevant spans of text, and systematically categorized the spans in terms of a coding scheme. The SPH’s coding scheme captured information about agencies and actors in a state’s public health system who were directed by statute to interact with one another in particular ways while dealing with public health emergencies. Based on the coded information, the SPH constructed statutory network diagrams of legally mandated interactions among actors. These network diagrams provide insight into those statutory texts that directed the interactions.

There are a variety of different approaches to representing legal information in order to make that information usable in legal analytics or automated reasoning systems. This variety of approaches stems from the variety of tasks that legal practitioners wish to undertake, and there is no single representation technique that squarely fits all tasks and contexts. This chapter will survey the methods available for representing legal information, and the benefits that each method can bring to the tools used in legal work. This section will also provide a high-level overview of approaches that have been developed to represent laws expressed as statutes, rules, and regulations. It will also survey the standards that have been developed and that are emerging for representing legal information. An overview of methods for representing case law will also be provided, along with a summary of how interpretation of the law can be achieved through automated reasoning mechanisms based on computational models of argument. This chapter will also consider methods of conceptualizing and reasoning about the relations within and between legal documents such as contracts, along with techniques to represent wider networks of information, such as document citations.

Every once in a while, a ready-to-use data set falls down the chimney like a diamond in a gift box, perfectly suited to the problem at hand. Unfortunately, what we usually have is a pile of coal and a rusty shovel. This is because most data resides in unstructured and disparate information systems or data sources. In order to apply most informatics methods, including a markup system like XML, we must first retrieve and then preprocess data from these sources to produce a structured, linked data set. These phases, sometimes colloquially referred to as data scraping, cleaning, wrangling, or “munging,” are arguably more important, and typically more time-consuming, than many other tasks in legal informatics.

Gamification refers to the use of game mechanics merged with behavioral analytics in a non-game setting.1 Gamification is used to improve production and performance in the workplace by engaging the user to behave in a way that is aligned with the goals of the business. Gamification occurs when a process, such as entering billable hours into the firm’s software or filling out an online client intake form, is mixed with game elements in such a way that firm members are motivated to complete tasks in a more desirable way. Businesses have used gamification strategies, with differing levels of sophistication, on issues including customer relationship management, training, market research, business intelligence, and education. Other professions, many in health care, are now also turning to gamification to increase engagement in a number of workplace processes for both their staff, and the clients they serve.

As highlighted in , the field of artificial intelligence (AI) is broad and embraces a wide range of approaches. While it is of course possible to imagine the linkages between topics such as machine vision and robotics to the business, practice, and delivery of law, the two most relevant topics within the AI landscape are natural language processing (NLP) (discussed in Chapter 2.8) and machine learning (ML).

Big Law has been described as being in the throes of a painful transformation brought about by factors such as globalization, the increased use of technology, and a transition from a supply-driven market to a demand-driven one. A common framework for such upheaval is Clayton Christensen’s The Innovator’s Dilemma, which generally portends an inevitable collapse of market incumbents when they cater to the performance requirements of their high-value customers’ demands. Big Law is not immune to the principles of The Innovator’s Dilemma. However, neither the disruptive nor sustaining innovation described in Christensen’s work seem to adequately characterize the changes occurring. In this chapter, we describe a hybrid model, adaptive innovation, that takes into account the opposing forces in play. As with most other sectors, lawyers have argued that Big Law is different. This chapter reviews some of the most cited factors predicting and denying the demise of Big Law. We argue that market-imposed values such as quality, efficiency, and ROI will likely dominate over reputation and comprehensiveness, forcing a fundamental change in many common features of Big Law. However, law firms will likely remain an inevitable mechanism for the delivery of services, albeit under a different model.