Legal services providers have long looked toward technology as a means to define value propositions across their products, processes, and business models. Consider the impact technology like e-billing had on attorney timekeeping or the use of Lexis and Westlaw had on legal research. 

Artificial intelligence (AI) represents the latest wave of technology shaping—and defining—the way consumers view products and how organizations deliver services. AI infiltrates almost every aspect of our world today, revealing itself in autonomous vehicles, app-based health care consultations, smart building capabilities, and various platforms like Netflix that predict viewer preferences with unnerving accuracy. 

Retail, health care, finance, manufacturing, and even legal services are all fields that have integrated AI applications into their business models. The Coca-Cola Company, serves as an exemplar to this observation. Coca-Cola still produces the classic Coke soft drinks consumed by billions each day but in 2017, Coca-Cola successfully launched Cherry Sprite based on data it had generated from consumer behavior and preferences at its self-serve soda kiosks. More than 130 years after its inception, the company remains a pivotal force in the food and beverage industry because of its commitment to leveraging and acting on data largely derived through artificial intelligence. 

AI holds equal promise for legal services. The potential for AI to automate manual tasks and help clients solve some of their greatest business challenges is vast and untapped. 

Machine Learning

Machine learning mimics critical knowledge functions and then automates time consuming and laborious tasks such as legal research and document review. Consequently, attorneys and providers have freedom to focus on more advanced functions, such how research and documents apply to their cases. 

Powered by data and effectuated through human training and interaction, artificial intelligence enables algorithms to be applied to data for the purpose of analysis, pattern identification, and response automation in an application. The entire process combined is known as 'machine learning'. 

The legal industry first commercially applied machine learning with the emergence of technology enabled legal services. A 2012 opinion issued by Judge Andrew Peck in the matter of DaSilva Moore v. Publicis Groupe et. al. was the first to endorse the use of computer-assistance in the analysis of electronically stored information (ESI) in support of e-discovery under the Federal Rules of Civil Procedure. Since then, various state, federal, and high courts have validated the appropriateness of machine learning and AI in response to document requests and predictive coding. 

Artificial Intelligence in Practice

Machine learning algorithms in data analysis is a well-established practice to help clients maximize efficiencies and reduce costs. The types of algorithms that are best suited for legal services, like support vector machines and logistic regression, enables the identification and meaningful classification of data within relatively small samples. When implemented across the spectrum of business and legal services, machine learning creates process efficiencies that drive decisive and measurable value. 

Broadly, two categories of legal practice that AI modulates and influences can serve as exemplars for the application of machine learning: all points in the electronic discovery reference model (EDRM) and case research. 

Electronic Discovery Reference Model Applications 

The foundation of modern-day e-discovery, continuous active learning (CAL) and supporting technology-assisted review (TAR) are powered through machine learning algorithms. All along the Electronic Discovery Reference Model (EDRM), machine learning supports each step. Beginning with the first several stages of identification, preservation, and collection, a subject matter expert, or team of experts, collects a data sample, analyzes and codes it, and feeds the results to the AI-powered machine, which draws conclusions about relevance and responsiveness of documents within the full data set. The result is increased efficiency and accuracy that has redefined the discovery process. 

Into the processing and review stage in the EDRM, handling files and reviewing critical data has traditionally been a labor-intensive component. Indeed, document review can account for the majority of e-discovery costs in any particular matter. With the proliferation of data combined with the commoditization of manual labor to support document review, machine-learning tools like TAR have become the greatest vehicles for cost savings and efficiency. 

The accessibility of voluminous structured and unstructured corporate and legal data remains a prevalent concern for corporate governance leaders, legal counsel, and risk managers. AI software that intelligently organizes and synthesizes data may help companies determine the value of its data, whether it is subject to regulatory restrictions, or legal holds, as well as utilize insights for business intelligence, compliance, or litigation. 

Case Research 

Machine learning tools enable practitioners to access to the latest, most relevant case law, decisions, and courtroom trends influencing their practice and their clients' matters in real time. Within a cloud-based infrastructure, machine learning permits searches that pull data from various operational systems simultaneously then returning the results in minutes. Recent software developments now support natural language searches. The advancement will return search results that are more comprehensive and capture critical information. Furthermore, the collection and analysis of historical data through artificial intelligence and applied machine learning allows attorneys to forecast likely outcomes, evaluating with precision how their cases will perform under certain conditions. 

The Power of AI Unleashed

As applications of artificial intelligence and machine learning proliferate, and the efficiencies they promote become more tangible, clients will come to expect their broadened application to even more legal services and functions. As far as the legal industry has come with the integration of artificial intelligence and machine learning, the greatest potential has not yet been unlocked. 

Historically, practitioners have accepted the imperfect nature of document review processes. Machine learning tools have remarkably enhanced the quality of document review, particularly as it applies to identifying document relevance and responsiveness. Yet, while years of judicial opinion exists in support of the application of machine learning to identify document relevance, use of advanced learning functions lag. The technology exists to not only determine a document's relevance, but to extract data from those deemed relevant, such as whether they contain privileged or confidential information that must be redacted, which may help drive better client outcomes. 

Applied machine learning represents a solution to accuracy and consistency concerns in document review and corporate governance. Beyond the applicability of machine learning to identify document relevance within a specific matter, there exists enormous potential to automate the analysis and decision-making that occurs post-review. Machine algorithms applied to culled data sets may formulate more accurate, consistent decisions about documents than could be achieved by human analysts. 

Most machine learning automates and supplements specific manual tasks. In contrast to machine learning is deep learning, which signifies the application of machine learning to broader data representations in supervised, unsupervised, or hybrid alternatives. Deep learning is what drives autonomous vehicles and builds champion chess players, and it holds significant potential for legal services. For example, when e-discovery technologies are deployed in offensive strategies, deep learning may empower advanced entity normalization functions capable of collecting and filtering data about people, places, and events and, subsequently, analyzing intent, opportunity, sentiment, and rationalization. The derivatives of deep learning may help legal services providers better understand risks and evaluate what legal arguments will hold weight. 

Artificial intelligence will be the chief technological influence of tomorrow's legal services models. AI, together with automation and machine learning, will give smaller practices broader shoulders to compete with their established predecessors on a more even playing field. Firms whose workflow intersects with artificial intelligence at every turn can scale their services at economically competitive rates, creating differentiation and driving value. As the evolution of technology continues, the potential for AI to revolutionize the legal industry is limitless. 

Jon Lavinder is an Epiq Expert who leads organization's practice on the use of analytics and AI to improve legal discovery outcomes for clients including many fortune 50 corporations. Jon and his team have guided clients to positive outcomes using analytics and AI workflow for thousands of legal discovery projects over the past decade.