Each day, millions of individuals around the world experience frustration, bewilderment, and failure when trying to understand or assert their legal rights. Without the assistance of an attorney, survivors struggle to obtain simple but necessary protection orders; tenants face horrific housing conditions or often illegal self-help evictions; consumers are victimized by assembly-line civil suits; and the retired, injured, or disabled struggle to obtain the benefits to which they are entitled.

Equality under the law is a fundamental principle of justice,Footnote ¹ but it is increasingly elusive for most low-income and even moderate-income individuals throughout the world. Few individuals have the legal training and knowledge needed to effectively represent themselves in court, even in simple civil cases, or to navigate any but the most straightforward government benefits processes. Adequate legal representation imposes financial burdens that few individuals can afford. The growth in the number of individuals who lack the resources for adequate legal representation has been characterized as an “access-to-justice crisis.”Footnote ²

Ironically, the access to justice crisis has coincided with dramatic improvements in artificial intelligence (“AI”) capabilities applicable to law. While the development of techniques for automating legal problem solving has been an area of active research for decades, their practical applications have primarily targeted attorneys and law firms, since they have had the greatest financial resources and incentives to adopt new legal technologies. However, for many years, a little-known community of judges, academics, and legal-aid attorneys has been exploring technologies to assist unrepresented litigants. The rapid advancement of AI techniques appropriate for law, particularly language processing, machine learning, and generative AI, has created an opportunity to extend this pioneering work to dramatically equalize access to justice.

The focus of this book is on technologies to help unrepresented litigants access an alternative source for the knowledge and skills that attorneys typically offer. Unrepresented litigants are individuals who, by circumstance or choice, represent themselves in legal proceedings without the assistance of a lawyer. Their unrepresented status may be due to various factors, including the inability to afford legal representation, lack of access to legal services, or the desire to handle their cases independently. Unrepresented litigants navigate the complexities of the legal system, which is built for lawyers and legal professionals, on their own, undertaking tasks such as filing documents, preparing legal arguments, and presenting their cases in courts. An unrepresented litigant is also referred to as a “self-represented litigant” (e.g., US, Canada, Australia), “litigant in person” (e.g., India, Singapore, UK), and “pro se litigant” (e.g., US).Footnote ³

AI systems that aim to improve the ability of users to achieve their goals are called decision support systems or intelligent assistants, and when the goals are law-related, such systems are termed legal AI assistants. AI-driven tools can mitigate the problems unrepresented people face by streamlining legal processes and improving the efficiency of legal services.

Equalizing Justice is a resource for those who want to develop, use, or understand legal AI assistants as tools for improving access to justice. We focus on technologies that automate tasks faced by unrepresented litigants navigating the legal system. Our approach encompasses not only generative AI and machine learning but also more traditional computational logic systems. From document automation and computational logic to interactive technologies like dialogue systems and case elicitation tools, we examine a spectrum of tools that can meaningfully assist unrepresented litigants. By considering both advanced AI applications and more conventional computational approaches, we analyze how these technologies can collectively reduce barriers to justice by simplifying complex legal processes, improving information accessibility, and enhancing the overall experience for those without legal representation.

This book does not address applications of AI or other computer technology that function as autonomous decision processes, such as systems for calculating the probability of recidivism for the purpose of recommending sentences,Footnote ⁴ employment and tenant screening systems,Footnote ⁵ or systems for automatically calculating benefits.Footnote ⁶ Autonomous systems that impact individual rights and enforce penalties raise a range of procedural due process issues, including how such systems “do not provide any opportunity for meaningful cross-examination, knowledge of opposing evidence, or the true reasoning behind a decision.”Footnote ⁷ Examination of issues regarding autonomous systems is an important topic, but it falls outside the scope of this book, which focuses on legal AI assistants to inform and empower the decisions of unrepresented litigants.

We argue that acceptance and adoption of legal AI technology in response to this access to justice crisis are imperative for two reasons. The first is simple fairness. A legal system that the average person cannot understand or use effectively fails to satisfy the fundamental principle of equality under the law. AI techniques that enable unrepresented individuals to understand or navigate the system governing their legal rights would help reduce the disparity in access to justice, thereby helping to level the playing field for all litigants, regardless of their resources.

A second reason is that a lack of access to justice gives rise to needless social costs. AI systems that handle common legal tasks can eliminate redundant effort across the legal system, where countless individuals separately struggle to understand identical procedures and requirements. A single investment in a system to automate the rules and procedures for common legal situations could relieve this redundancy. From this perspective, legal AI assistants for unrepresented litigants can be viewed as public goods, similar to roads, schools, and courthouses, whose costs and benefits are shared by all citizens. While there has long been a consensus on the value of publicly funded courts and judges, rather than passing the cost to litigants, the idea of treating software to improve access to justice as a similarly valued public good is relatively new.

Part I, Addressing the Crisis in Access to Justice with Legal AI, opens with Chapter 1, where we examine the magnitude and costs of the global access to justice crisis. Chapter 2 analyzes the language gap and explores the potential of generative AI to bridge it. Chapter 3 investigates the needs of unrepresented litigants and the tasks they must accomplish to assert their rights and defenses, laying the groundwork for future chapters to explain how AI and other technologies can assist them.

Part II, Automating Legal Tasks with AI and Other Technologies, demonstrates how certain legal AI and other technologies can automate the tasks unrepresented litigants need to assert their rights and engage with the legal system. Chapter 4 introduces generative AI, highlighting the recent technology that offers unparalleled capabilities for bridging the language gap, while also presenting many nuances and pitfalls. Chapter 5 traces the history of legal AI and its evolution. Chapter 6, on computational logic, explores techniques for modeling the rule-governed nature of law. Chapter 7, on machine learning, demonstrates capabilities for predicting case decisions and generalizing over cases. Chapter 8, on document automation, covers the widely adopted legal technology. Chapter 9, on interactive legal technologies, discusses chatbots, case elicitation, and legal design.

Part III, Promising Applications, highlights how legal AI and other technologies can address the needs of civil litigants, as discussed in Chapter 3. Chapter 10 emphasizes the importance of integrating these technologies to solve the full spectrum of litigants’ needs. Chapter 11 explores online dispute resolution (“ODR”), a promising technological approach for equalizing the bias in favor of large-scale institutional plaintiffs. Chapter 12 reviews some uses of large language models (“LLMs”) for the specific goal of increasing access to justice.

Part IV, Today’s Access to Justice Systems, profiles tools offered both for free and at a cost. Chapter 13 profiles tools that are hosted directly by courts and tribunals, law schools, legal aid organizations, nonprofits, and companies. It also acknowledges that a business model for lawyers to provide low-cost consumer-facing tools and legal services to those in need is a critical part of the solution, highlighting several of the companies that do so already.

Part V, Barriers and Risks, acknowledges that while AI systems have great potential to benefit unrepresented litigants, there are also important barriers and risks to consider. Chapter 14 tackles prohibitions against the unauthorized practice of law (“UPL”), one of the most significant institutional barriers to adopting legal AI assistants. Chapter 15, on AI assurance, underscores the importance of reliability, privacy and accountability in legal AI technology.

Part VI, Recommendations, offers both technological and institutional recommendations for developers, legal aid organizations, nonprofits, courts, and other institutions looking to leverage AI or other technologies to assist unrepresented litigants. Chapter 16 outlines technological recommendations, specifically technical methods for evaluating AI applications for legal tasks. Chapter 17 addresses institutional recommendations and emphasizes that technological innovations alone will not be sufficient to solve the access to justice crisis. A variety of institutional reforms are needed, including simplifying legal procedures, increasing funding for legal services, and reducing restrictions on UPL, which all restrict the availability and affordability of legal services for unrepresented litigants.