TAR + Advanced AI: The Future is Now

Issue link:

Contents of this Issue


Page 5 of 12

WHITE PAPER: TAR + ADVANCED AI – THE FUTURE IS NOW | 6 applications within them, etc.) rather than communicating and working within the limited number of applications that were available 10 years ago (think: databases, email, Microsoft Word and Excel, etc.). To have true efficacy in this modern environment, systems trained to classify data must be able to handle the additional volume, as well as grasp the different ways information is communicated within constantly evolving data sources. This requires a combination of technologies, heavily leveraging the more advanced subsets of AI available today – specifically, the subsets of deep learning and natural language processing (NLP), both of which are explained in more depth later in this white paper. Like the original ediscovery and TAR evolutions that came before, if attorneys are not willing to evolve and adapt with technology to manage the realities of modern data volumes and diversification, they risk falling down on their ediscovery requirements and subsequently failing their clients. Accordingly, we believe that updated TAR processes that utilize advanced AI will soon not only be considered an optional tool for forward-thinking attorneys who want to provide clients with better and more efficient discovery processes, but will also be an ethical duty on the part of attorneys necessary to meet big data ediscovery requirements. What is advanced AI? As previously noted, the term "AI" in ediscovery has become muddied – somehow simultaneously bringing to mind the somewhat limited functionality of TAR while conjuring futuristic images of human-like robots and fully self-driving cars. Because AI encompasses a large swath of technology, simply using the term "AI" can result in conflating completely different technologies. Thus, this section will begin by defining advanced AI as it applies to ediscovery and how that technology can be used to improve upon ediscovery – and more specifically TAR 1.0 and TAR 2.0 workflows – in today's more complicated and voluminous datasets. Broadly speaking, AI refers to the "science and engineering of making intelligent machines, especially intelligent computer programs." 15 It encompasses the subfields of machine learning and deep learning. 16 Machine learning "focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy where those algorithms are trained to make classifications or predictions." 17 This is especially important in ediscovery, as data volumes continue to grow and diversify. Older subsets of machine learning, like supervised machine learning used in TAR, require human intervention to process data. 18 Deep learning, while a subset of machine learning, "automates much of the feature extraction piece of the process, eliminating some of the manual human intervention required and enabling the use of larger datasets." 19 Natural language processing (NLP) is also a separate branch of machine learning that involves training computers to understand text and spoken word in the same way that humans understand it. It combines rule-based modeling of the human language with statistical, machine learning, and deep learning models to process human language and "understand" its full meaning, including the intent and sentiment of the writer (or speaker). 20 Thus, when we refer to "advanced AI" in this paper, we are referring to a combination of the aforementioned technology that utilizes multiple branches of AI (i.e., traditional machine learning in combination with deep learning and NLP) to make more accurate and efficient predictions within modern datasets (i.e., datasets that are more voluminous and diverse than even five years ago). Below, we've defined a few other technology terms that will be helpful as we delve further into the topic of "advanced AI". AI TERMS TO KNOW Artificial Intelligence (AI): Broadly, AI refers to the "science and engineering of making intelligent machines, especially intelligent computer programs." 21 IBM clarifies that AI is "a field, which combines computer science and robust datasets, to enable problem-solving." 22 AI also encompasses subfields of machine learning and deep learning. 23 Big Data: Datasets that are beyond the ability of traditional relational databases to capture, manage, and process the data with low latency due to high volume, high velocity, and/or high variety. These datasets are more complex than traditional datasets due to a variety of modern-day influences, including AI, mobile devices, social media, etc. 24

Articles in this issue

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