An Ethical Artificial Intelligence: LegalTech in Southern California

Many are wary of artificial intelligence (AI). Even Elon Musk, the business magnet behind the push for autonomous vehicles, has expressed concerns about a world in which machines outpace humans. Musk is not alone. Legal practitioners are equally skeptical of the proliferation of artificial intelligence. Some point to their deployment in the criminal justice system. Courts and corrections departments across the United States now use algorithms to help determine a defendant’s risk of committing another crime or failing to appear for a court hearing. These algorithmic outputs (purchased from private businesses) inform decisions about bail, sentencing, and parole, raising concerns about the black-boxed nature of proprietary softwares as well as the hidden biases programmed into these technologies.

Why the Push for Legal Tech? 

As the first step to artificial intelligence, machine learning involves “the recognition of patterns and the automation of processes based on human input.” Legal tech innovators have started to apply this same procedure to document review, discovery, and day-to-day legal research activities. “It may even be considered legal malpractice not to use AI,” says Tom Girardi , a civil litigator based in Los Angeles County. “It would be analogous to a lawyer in the late twentieth century still doing everything by hand when this person could use a computer.” Girardi’s comments emerge within a context of extraordinarily high legal costs. According to Leah Wilson , Executive Director of the State Bar of California, “too many Californians needing legal services cannot afford an attorney.” Many expect AI-powered legal technologies will be able to help bridge this gap in access. “It’s a lawyer’s job to solve a problem as quickly and inexpensively as possible,” Girardi begins. “AI will be a godsend because it’ll give lawyers the information they need to resolve conflicts faster.” 

Creating an Accessible and Transparent Legal System 

Andrew Arruda is an attorney whose legal research platform, ROSS Intelligence, was developed to help ensure that “everyone across the state [of California] can gain access to high-quality, affordable legal services.” ROSS changes the ways in which attorneys conduct legal research, providing features that enable attorneys to quickly identify any weaknesses in the arguments used by opposing counsel. Without a cohort of junior associates at which to throw legal research tasks, small law firms and sole practitioners (as well as their clients) are poised to benefit the most from AI-powered legal technologies. In fact, one study conducted by Duke University, the University of Southern California, and Stanford University demonstrated that artificial intelligence can outpace experienced lawyers when it comes to reviewing legal documents. During this experiment, the AI software achieved a 95 percent accuracy rate when it reviewed a contract in 26 seconds. The human lawyers, on the other hand, took an average of 92 minutes to achieve an 85 percent accuracy rate. Consider another example. Trellis Research has compiled an archive of state trial court rulings for California, Florida, New York, and Texas. Using this archive, Trellis has developed an analytics platform that brings together the unstructured data from counties across each state and restructures the information in ways that allow litigators to draw a plethora of strategic insights. Earlier this year, one firm based in Southern California utilized this platform to develop a custom analytics strategy report on settlement amounts in class action cases with PAGA claims. The report included detailed information about the settlement amounts and attorneys’ fees that judges had previously approved. The report then compared these amounts with the number of weeks for each trial. With this kind of analysis, the firm developed a settlement strategy that catered to the specific needs of its client. The firm could perform a cost-benefit analysis that would help it determine the number of weeks their client should continue with a trial before it would be more financially beneficial to settle. With these examples, we can begin to see how AI-powered legal technologies are democratizing access to the law. By making legal data more accessible, they even the playing field between large and small firms, all while reducing the costs of legal services for clients and lifting the veils that cloak case strategy from clients. 

Opening the Doors to Justice

Legal practitioners should continue to remain skeptical of legal technologies and their unchecked deployment across the industry. That skepticism, however, should also be accompanied by efforts to reimagine how these new technologies might be used to bring greater accessibility and transparency to our legal system. The work performed by ROSS Intelligence, Trellis Research, and other AI-powered legal tech providers suggest that artificial intelligence has the ability to not only expand who has access to legal protections, but it also has the ability to affect how we might harness that access in productive and equitable ways.

Prior to founding Trellis, Nicole Clark was a business litigation and labor and employment attorney who handled litigation in both state and federal courts. She regularly represented multinational corporations in claims ranging from high-profile trade secret disputes to complex class-action litigation. Frustrated by sending internal emails and collecting anecdotes on judges in order to make strategic case recommendations, she built Trellis to solve her own need for access to data, information, and analytics at the state trial court level. Prior to law school, Nicole attended Bard College, beginning her college coursework at the age of sixteen. She graduated with honors from University of Massachusetts Amherst with a BA in Journalism, and received her Juris Doctorate from Rutgers School of Law in Newark, NJ. Nicole sat for the Bar Exam in California, Massachusetts, and Connecticut, and remains licensed to practice law in all three states.