The Growth of Document Review AI
by Brandon Jack
Editor’s note: On February 11, 2019, the New York Times published an article about a recent study in which the authors reported that they had built an AI system for the diagnosis of common childhood illnesses.1 In 2018, radiologists learned of a Chinese contest which pitted an AI system against a team of 15 “senior doctors” in two rounds of competitions diagnosing brain tumors and predicting the expansion of brain hematomas. The AI system “won” handily.2 Some articles predicted the future demise of humans in radiology; others argued that humans will never be replaced. An article in Harvard Business Review gave sage advice – just as apt to lawyers as to radiologists: These professionals “will need to adopt new skills and work processes … The only radiologists whose jobs will be threatened are the ones who refuse to work with AI.”3
One night last June, at Kensington Palace in London, “Luminance” was awarded “Best AI Start-Up” at The AIconics 2018. The judges selected Luminance for, among other things, its innovative approach to contract review. The delighted Luminace CEO, Emily Foges, crowed that “Luminance is truly revolutioni[z]ng contract review processes around the world.” Will Sacramento lawyers rush out and purchase Luminance? Not likely, but neither should lawyers ignore that Luminance and its competitors here.
Saving time, money, and providing top service to clients are the three powerful driving forces in the legal AI market. Since the inception of computers, word processors, and easily stored electronic data into the public marketplace, lawyers have had to deal with an ever-increasing volume of information.
The Luminance AI system reads, understands, analyzes, and compares vast quantities of documentation at a rapid rate while also learning from the interaction between the user and those documents. These tools, Luminace argues, permit the lawyers to focus their efforts on more important tasks that require a lawyer rather than a machine. Luminance has systems such as pattern recognition algorithms and unsupervised machine learning. The pattern recognition algorithms help the lawyer review large quantities of repetitious documents such as employment contracts, lease agreements, and purchase agreements. The software’s ability to compare and contrast increases the accuracy and efficiency of the lawyer because it can read a multitude of documents at the same time and will never get tired or miss any important differences in documents like a tired lawyer might. Luminance’s machine learning is allegedly unique among the legal AI market – for now – because it offers “unsupervised” machine learning. Unsupervised machine learning means that the software does not need any input from a human being to identify key similarities and differences across large numbers of documents because the software learns from the user’s interactions with the documents in real time. Luminance claims to be apart from other legal AI or big data analysis algorithms in this respect because other systems usually require the user to directly supervise the machine learning algorithm for even the most basic functions.
No AI roll out would be complete without addressing the fear that AI will eliminate jobs for many qualified people. No, says Luminance CEO Foges, “the legal industry should not be worried about artificial intelligence.” The advent of legal AI “is not about robots taking lawyers’ jobs; this is about technology helping lawyers to do a better job.” Foges argues that the AI platform is simply a tool that aids and will not replace the lawyer because machines cannot conduct most of the functions that are required in being a lawyer. Time will tell whether this is true.