Artificial Intelligence in Legal Research

The recent literature and numerous social media posts have noted that artificial intelligence (AI) is fast becoming part of the legal practice landscape, including legal research. Let’s briefly summarize a few of these products that focus on legal research.

  1. CARA or Case Analysis Research Assistant is a tool developed by Casetext that can automatically review a document and look for cases or statutes that are relevant to the cited authority. By analyzing the document, CARA generates a list of additional authority that may be relevant to the cited references in the document.


  1. Judicata is a relatively new legal research search engine claiming to provide results that are precise, relevant, and simple. Judicata asserts it can do so because it has mapped the “legal genome” and developed a set of filters that allow search results to be narrowed to their core components. However, for now, Judicata only includes California law but eventually hopes to add all jurisdictions.


  1. ROSS uses IBM’s Watson, an artificial intelligence system capable of natural-language processing to filter through legal documents. Accordingly, “[i]nstead of searching for documents by keywords, one can ask questions in plain English . . .” The legal research robot is accessible via computer and billed as a subscription service. However, organizations providing legal assistance are given free access to the tool.


  1. KNOMOS is described by its CEO as leveraging “data visualization and machine learning to augment user experience and develop a connected knowledge network for legal information . . .” Users get an “instant overview of how multiple legal sources relate to one another, with enhanced discoverability of key results based on contextual data from other users . . .”


  1. Loom Analytics, a Canadian based system that “is data-driven legal research assistant that finds, classifies, and sorts case law . . .” Using a combination of legal analysis and machine learning, the system produces “hard numbers” on case law such as win/loss rates, judge ruling histories, litigation trends over time, and other like metrics.


  1. blueJ Legal uses IBM’s Watson computer to run it flagship product, Tax Foresight. According to its website, Tax Foresight collects and analyses the facts and findings of Canadian court cases to predict what a court would hold under different circumstances. The site claims that “[t]he information that Tax Foresight collects is sufficient for the system to reach the correct prediction in greater than 90% of the cases in out-of-sample testing.”

These are a few examples where AI has crept into the legal research process. The infusion of AI into many of today’s legal tasks will only increase over time.  But will it be enough to eliminate the research librarian or attorney? I think not. The unpredictability and random thought processes of humans are unparalleled and cannot be readily duplicated or replaced. There are simply an infinite number of variables, unknowns, and unpredictable scenarios that cannot be anticipated with mere algorithms.

%d bloggers like this: