In addition to having more data to manage, the expectation that companies will identify and protect sensitive data has also intensified, and the failure of any organization to maintain privacy protections could have devastating consequences.
Artificial intelligence (AI) and machine learning technologies have been used for years in support of discovery and litigation. But most in-house legal departments have other operational needs in addition to discovery and litigation, including general information governance activities, corporate investigations, data privacy and compliance, contract review, and the management of second requests to support mergers and acquisitions (M&A) and divestiture activities. With trends indicating a growing need for companies to seek more sophisticated and technologically advanced solutions, can AI be applied to address use cases in these other areas as well?
First at the AI Bat: Responsive Review
Since 2012, when New York Magistrate Judge Andrew J. Peck approved the use of technology-assisted review (TAR) in the Da Silva Moore case, the use of AI-based technologies has become increasingly common in support of electronic discovery in litigation. Today, lawyers and legal professionals commonly train supervised machine learning algorithms to streamline the labor-intensive task of document review in litigation, saving time and expense, while often improving the accuracy of review – if the work is done correctly.
Although workflows associated with TAR are by now well established, experience has shown that TAR is not a magic bullet. It still takes considerable time, effort and expertise to use TAR tools successfully, with human intelligence and intervention playing a significant role in the outcome. This is especially true when it comes to the inherent nuance of privilege review, which has essentially remained a manual effort. One hopeful sign is the increasing use of AI-based technologies for this function as well, which will accelerate the more laborious aspects of this time-consuming and costly process.
AI Beyond Document Review
It is important to keep in mind that most AI use cases in the legal realm involve the analysis of textual content. Legal challenges related to information governance, data privacy, compliance, investigations, contract review and M&A all have in common the need for analytics tools that interrogate massive volumes of text. The use of AI in these endeavors – whether machine learning algorithms, data classifiers or otherwise – generally requires both linguistic and analytics expertise behind the tool for a successful result. And, as opposed to a TAR workflow, which is typically applied to sets of documents that have previously been identified and collected, other potential legal use cases need to contend with data in place.
AI technology is already being used regularly to automate the review of day-to-day business contracts. With a typical Fortune 1,000 company maintaining 20,000 to 40,000 active contracts at any given time, this can result in huge time and cost savings, while also improving accuracy.
Below, we consider some of the other trends that corporate legal departments now face and explain why both linguistic and data analytics expertise are necessary parts of the AI skill set.
Information Governance: To say that corporate legal departments have more data to manage than ever before would be a significant understatement. According to StatInvestor, data volumes in the world have grown from 0.1 zettabytes in 2005 (zettabyte = 1 trillion gigabytes) to 47 zettabytes in 2020 – a growth of 470 times in just 15 years.
Part of the reason for such growth is the expanded variety of data sources that today’s corporations need to address. In addition to enterprise-wide systems that manage everything from accounting and customer relationships to email and work product generation, corporations need to account for data from mobile devices, social media and other cloud systems, collaboration apps like Slack and Microsoft Teams, and even (eventually) data from internet of things devices.
Aside from the costs associated with maintaining ever-growing data stores, the volume and complexity of this data presents a number of risks. With so much data to manage from so many different sources, much of which could (and likely should) be disposed of, the ability to analyze and assess data content has become a crucial corporate need. The challenge: How do you identify which data you can defensibly delete?
Any response, AI-enabled or otherwise, must be supported by human expertise.
Data Privacy: The growing number of data privacy laws comprise another trend impacting legal operations today. With many states and countries taking steps to require the protection of personally identifiable information (PII), companies need to know what personal information they have in their possession and where that information is stored. In May 2018, Europe implemented the General Data Protection Regulation to protect the data privacy rights of European individuals, and California implemented the California Consumer Privacy Act in January 2020 to do the same for California residents. (More resources can be found in the International Privacy Law Library.) Also, corporations are having to implement new workflows to respond to data subject access requests – i.e., requests from individuals about the way companies handle their personal data.
So, in addition to having more data to manage, the expectation that companies will identify and protect sensitive data has also intensified, and the failure of any organization to maintain privacy protections could have devastating consequences. The challenge: How do you identify data that needs to be protected?
Corporate Investigations: Corporate legal is also having to address an increase in corporate investigations. In the 2019 Corporate Investigations Survey (conducted by H5 and Above the Law), nearly two-thirds (63 percent) of respondents involved with corporate investigations expected the number of investigations to increase over the next three years. Since the pandemic, those concerns have only increased. For example, in a September 2020 survey conducted by the Association of Certified Fraud Examiners, 74 percent of surveyed certified fraud examiners indicated that preventing, detecting and investigating fraud in the COVID-19 era has been even more challenging than it was before the pandemic. The challenge: How do you find potential evidence or indicia of criminal activity within massive data stores?
Mergers, Acquisitions and Divestitures: Corporate investment activities such as mergers and acquisitions are keeping corporate legal departments busier than ever. According to Deloitte’s M&A trends 2020 survey, 63 percent of respondents expected M&A transaction activity to increase this year, with divestitures (driven by organizations seeking to cash in on high valuations and those aiming to reposition assets in advance of a downturn) potentially becoming more active. Those transactions may require closer scrutiny by the Federal Trade Commission for antitrust concerns, and corporations may be required to respond to a second request to provide more information about the transaction. These second requests often entail accelerated time frames. The challenge: How do you quickly identify responsive information for an impending HSR request?
Addressing the Challenges
It should be clear from the trends and challenges identified above that evolving use cases throughout the enterprise cry out for a sophisticated technological response. Notably, because they are primarily driven by the requirement to interrogate text, AI in the form of machine learning algorithms, natural language processing, and data classification methodologies can play a major role. Equally notable is that each challenge has a unique set of requirements to address. The application of AI solutions should be an iterative, methodical and scalable exercise. Any response, AI-enabled or otherwise, must be supported by human expertise – both in the analytical assessment of the requirements and the development of a solution. The maturity of the organization to handle advanced technologies is of no small consequence either, and the care and feeding of such systems is an ongoing requirement.
The consultancy firm McKinsey estimates that 22 percent of a lawyer’s job and 35 percent of a law clerk or paralegal’s job can be automated, enabling a focus on more vital and strategic work. There are plenty of opportunities to leverage AI technology to support that automation effort. The good news is that development of AI tools to use within the enterprise is a rapidly developing field. Without a solid foundation upon which to build AI capabilities, however, ad hoc applications of AI tools may fail to meet expectations. What is required is a strategic response that includes a prioritized assessment of the use cases to be addressed and an expert analysis of the best AI or advanced analytics tools for each use case. More good news is that the expertise to provide both the assessment and the potential solution is readily available in the legal services ecosystem.
Published February 12, 2021.