Skip to main content

Artificial Intelligence and the Courts: Materials for Judges

Project Scope

With the financial support of the National Institute of Standards and Technology (NIST), AAAS has undertaken a one-year project aimed at developing preliminary information resources necessary to support judges as they address an increasing number of cases involving artificial intelligence (AI).

Accordingly, materials for judges have been generated that can assist them to become familiar with various aspects of artificial intelligence (AI), which is increasingly used in the legal field and beyond.  Indeed, courts likely will have to handle more and more cases in which AI is central to the matter, and AAAS hopes that the written and audio materials it is making available cogently provide accurate information to help judges in their adjudication of such cases. 

To that end, the materials generated by this project are primarily individual papers, prepared by experts in the relevant field, and finalized through a process that ensures both the technical accuracy of the content and its utility for judges.  The materials also include podcasts, and accompanying transcripts thereof, in which experts discuss topics that should help put the increasing use of AI in the legal field in context, raise questions that judges should bear in mind as this phenomenon advances, and offer some suggestions for judges to consider.

In addition to making these materials available via AAAS’ website, AAAS intends to have select partner organizations, that work with state and federal judicial education institutions, make the materials freely available to judges through their existing online databases and other digital repositories.

The production of these materials, as part of the Artificial Intelligence and the Courts: Materials for Judges series, was funded by the NIST through Award 60NANB21D031.

Disclaimer: The opinions, findings, and conclusions or recommendations expressed in these materials do not necessarily reflect the views of the AAAS Board of Directors, its Council and membership, or NIST.

Reference Materials for Judges

Teal background, white text, "Artificial Intelligence and the Courts: Materials for Judges - Foundational Issues and Glossary."

Foundational Issues and Glossary

This paper provides a very inclusive introduction to numerous key concepts with which judges and court personnel may need to be familiar. These include: The elements and variety of artificial intelligence (AI) systems; how these may be designed, developed, or deployed; and key issues with respect to the limits and risks associated with AI.

For convenient reference, the Glossary provides, in alphabetical order, a list of important terms and words (from, e.g., “Acoustical Processing,” to “GAN – General Adversarial Networks,” to “Responsible AI,” to “VR – Virtual Reality”) and their definitions. As warranted, differences in nuances or usage are noted. Additionally, certain common abbreviations or acronyms are included.

Red background, white text, "Artificial Intelligence and the Courts: Materials for Judges - AI, Trustworthiness, and Litigation."

Artificial Intelligence, Trustworthiness and Litigation

Although few court decisions have squarely addressed the admissibility of artificial intelligence (AI) evidence in proceedings governed by the Federal Rules of Evidence, or their state-law equivalents, this paper focuses on key considerations for the use of AI evidence in court cases. The paper defines the concept of “trustworthiness” as being the sum total of a number of interrelated requirements found within the rules of evidence that govern court cases.

Purple background, white text, "Artificial Intelligence and the Courts: Materials for Judges - AI, Legal Research, and Judicial Analytics."

Artificial Intelligence, Legal Research and Judicial Analytics

This paper touches on two matters—often overlooked—but of potentially very direct significance to judges: How artificially intelligent systems (AI) may affect the process and results of legal research, and whether or how assessments (i.e., “judicial analytics”) of judges’ prior rulings, decisions, or even style, might be leveraged by parties’ counsel to gain some advantage. That is, key-word searches, or other forms of “technology assisted review,” may be more or less effective, or be affected by, the way search-engines are designed. Separately, but increasingly, vendors offer products—based on analyses of judges’ behaviors and rulings—that purport to provide insights that will, in turn, reduce risks for litigants or parties. Finally, it may also be that AI will be able to usefully shed light on whether or when reforms to rules or procedures have proven to be affective in improving the administration of justice. 

Yellow background, white text, "Artificial Intelligence and the Courts: Materials for Judges - Artificial Intelligence and Bias - An Evaluation"

Artificial Intelligence and Bias - An Evaluation

As the authors of this paper underscore, “Judges should expect that problems related to bias are likely to manifest, in one form or another, across virtually every field where AI decision-making has become popularized.” Indeed, the problem-set of myriad ways bias can inhere in AI is of increasing interest and, to address it, a variety of perspectives is required; this paper cogently incorporates many important current viewpoints.* Courts will want to consider when or how bias relates to discriminatory impacts (e.g., in employment or housing cases), as well as the various possible origins of biases in a given AI system (e.g., datasets, design, or deployment). Also, difficulties with the “explainability” of AI may complicate showing the requisite discriminatory intent, but the “FAccT” (fairness, accountability and transparency) framework may help courts, as well as developers of AI systems, to understand the extent or nature of possible biases.

More Information