Marketing

TAR + Advanced AI: The Future is Now

Issue link: https://content.lighthouseglobal.com/i/1412483

Contents of this Issue

Navigation

Page 2 of 12

WHITE PAPER: TAR + ADVANCED AI – THE FUTURE IS NOW | 3 Introduction The concept of discovery in litigation dates to Common Law and the English Court of Chancery. 1 However, it wasn't until the late 1990s that the revolutionary rise of computing began to dramatically increase electronically stored information (ESI), catalyzing the first dramatic change to the civil discovery process in decades. The new document types forced lawyers and judges to change age-old discovery and litigation processes to reflect the reality of how people were now storing and exchanging information. By 2006, attorneys began to grapple with new requirements to preserve and produce ESI during discovery, due to groundbreaking amendments made to the Federal Rules of Civil Procedure (FRCP). Case teams faced with large volumes of ESI were forced to hire big teams of contract attorneys to individually review electronic documents, flagging important information and separately tagging the documents for responsiveness, privilege, confidentiality, etc. Depending on the volume, complexity, and number of attorneys, this process could take months or even years of work, billed out hourly. Enter more advanced technology. Beginning in the mid-to-late 2000s, case teams started to use a subset of artificial intelligence (AI) technology called "machine learning" to predict the responsiveness of documents, thereby expediting the review process. This process became known as technology assisted review (TAR). As the legal industry's first foray into AI, the term "TAR" was often conflated with the technology behind the tool – even though AI actually encompasses a much broader set of technology. Even today, when more advanced AI tools exist, many attorneys think of TAR technology when they hear the term "AI" used in the ediscovery context. Conversely, other attorneys may think of a myriad of unrelated ediscovery tools that do not, in fact, fall within the definition of AI. This muddiness around terminology has become a detriment to the important ediscovery work that attorneys perform today. Now more than ever, attorneys must have a solid grasp of how AI technology works to choose the best available tools and meet obligations to their clients and organizations. That's where we come in. We know not all attorneys are technologists. The goal of this paper is to deepen the reader's understanding of AI within the context of ediscovery and provide answers to common questions. We will focus on TAR to provide a familiar framework with which to talk about applications of advanced AI in ediscovery. In the remainder of this paper, we will outline the history and definition of TAR, define advanced AI, and then delve into the interconnection between TAR and advanced AI, how to choose the appropriate technology, and when to use it. What is TAR? To discuss how TAR can be revolutionized by combining it with the power of advanced AI, it is first necessary to level set by defining what TAR is and how it has evolved thus far. Currently, TAR is performed by machine learning algorithms that classify documents for responsiveness based on human input or "training." This classification allows attorneys to efficiently prioritize the most important documents for eyes-on review. Further, when agreed to by opposing counsel and/or a relevant governing body, the party performing TAR may also be able to avoid reviewing documents that the tool has determined are very unlikely to be responsive. In this way, TAR can significantly reduce the number of documents humans need to review, thus it has historically been helpful (and sometimes critical) when dealing with short deadlines or larger volumes of data. THE HISTORY OF TAR TAR as we know it today evolved from a more basic form of data classification performed by machine learning algorithms, which was first outlined in Anne Kershaw's 2005 study entitled "Automated Document Review Proves Its Reliability." 2 In that study, Ms. Kershaw's research found that "electronic relevance assessment application and process reduced the chances of missing relevant documents by more than 90 percent." 3 That research helped pave the way for more widespread acceptance of this type of technology in the ediscovery space.

Articles in this issue

view archives of Marketing - TAR + Advanced AI: The Future is Now