Recently thousands of New York attorneys met in New York City to attend the Annual Meeting of the Commercial and Federal Section (ComFed) of the New York State Bar Association. Of particular interest, ComFed litigators considered some of the ways technology is changing the practice of law and business, including the emerging role of AI in the legal system (and beyond).
Our discussion began with an overview of the technology that has permitted computers to do “intelligent” things, including completing the review and analysis of massive volumes of data in a very short period of time, and the profound impact such technological advancements has had on the legal profession.
With the advent of supervised machine learning in the past several years, lawyers and businesses have been introduced to programs that rely upon Technology Assisted Review (“TAR”) that, with the input of trained professionals, review and analyze data quickly in order to identify key documents and data to be further reviewed by trained professionals.
Specifically, TAR technology is defined as: “A process for prioritizing or categorizing an entire collection of documents using computer technologies that harness human judgments of one or more subject matter expert(s) on a small subset of the documents, and then extrapolate those judgments to the remaining documents in a collection.”
You might ask yourself, “Why do I care about TAR?” Here’s why: TAR technology is becoming more common in the business world amongst in-house and outside counsel as a tool to manage data-driven discovery. Further, TAR programs can be used to implement and monitor regulatory compliance and risk management programs. In short, TAR technology has the potential to save businesses money and headaches both inside and outside the context of litigation and is a budding technology concept that is worth exploring within the context of your business and within the context of litigation.
We also discussed how predictive analysis based upon a proprietary algorithm, COMPAS (Correctional Offender Management Profiling for Alternative Sanctions), is being used by many courts throughout the country in connection with sentencing persons convicted of crimes. The premise behind computer-generated recommendations is that the algorithm considers vast quantities of data, applies that data to the specific case at issue, and generates an assessment of the convicted person’s predicted risk of recidivism. However, because the program is proprietary, the specific information considered by COMPAS is not known by the general public, including those receiving the recommendations.
Similarly, predictive analytics about judges and the likely outcome of particular cases before trial is now readily available to lawyers and litigants. According to the providers, these predictions are based upon proprietary programs powered by natural language processing, data science, and machine learning.
Predictive analysis programs are available in other contexts as well. For example, predictive analysis algorithms have been used in designing programs for Human Resource Departments as a tool to eliminate bias in hiring personnel, amongst other things.
In theory, such predictive analysis is a means to provide an “unbiased” assessment of persons. In practice, though, any flaws in the algorithms used, flaws in the data sets, or flaws in the information provided for consideration would necessarily skew the results.
So, what is the takeaway here? Predictive data analytics/analysis is becoming more commonplace in courts and in the workplace. As technology continues to improve, its use will likely only grow. Thus, the users of such information have an increasing responsibility to monitor and assess the algorithms and data sets applied by such programs to ensure that they are meeting the goal of eliminating bias in the courts and workplace rather than inadvertently perpetuating it.