Robot lawyers have been getting so much attention lately that AI-and-law thought leaders believe we have reached peak hype. Journalists have responded by toning down their headlines to better manage expectations. For example, last month the New York Times ran an article titled, “A.I. Is Doing Legal Work. But It Won’t Replace Lawyers, Yet,” and the ABA Journal gently warned, “The robot lawyers are coming (to help, not to take your jobs).”
The Times article explains that automation generally happens task by task. So, even if AI can scan documents and predict which ones will be relevant to a legal case, other tasks such as actually advising a client or appearing in court cannot currently be performed by computers.
But for readers who are not well-versed in the law, these articles fail to answer some more foundational questions: What is legal research, anyway? And, if a computer can do the research, why would I still need a human lawyer?
To answer those questions, let’s look at a specific technology as an example. The company Casetext recently unveiled a tool called CARA to help lawyers do legal research. CARA stands for Case Analysis Research Assistant (it also means friend in Irish). This rollout coincides with Casetext’s announcement that it has secured $12 million in Series B funding, which will be used in part to further develop their AI capabilities.
How does CARA work? The user uploads a legal brief to Casetext’s website; CARA scans the brief and instantly returns a list of relevant cases that the brief failed to cite. I tested it out with a brief I wrote a few years ago when I was still practicing law. CARA’s speed and accuracy are truly astounding. It would have taken me hours of research to come up with the list of cases that CARA generated instantaneously.
But for those who have not experienced the drudgery of legal research firsthand, it’s hard to understand what this all means. Readers may be wondering, What is a “relevant case” and why is it so important to make sure you didn’t miss one?
See also: Bosch and NVIDIA create a supercomputer for self-driving AI
To appreciate the impact of tools like CARA, it’s important to have an understanding of how our legal system works. (Even if this was covered in a civics class at some point, most of us could use a refresher.) When a dispute ends up in court, the judge writes a decision resolving the case. Courts publish these decisions and they are collectively referred to as “case law.”
Our legal system is based on the principle of stare decisis, a Latin term meaning that cases should be decided consistently so that similar situations will yield similar results. Accordingly, when a dispute ends up in court, the lawyers and judge involved in the case look to older case law to see if the issue has been decided before. If it has, the older case will act as a precedent and the judge will follow its reasoning in deciding the current dispute.
Or, one of the lawyers may argue that the current situation is different enough from the older case to justify a different result. Even if there is a statute or regulation that seems to directly address the subject matter of a dispute, there may still be case law interpreting the language of the statute or regulation—filling in gaps or explaining how that rule applies to specific situation. In other words, no matter what type of dispute you have, it’s important to search all of the case law to see what judges have said about similar disputes in the past.
Enormous potential time savings
Before computers, cases were published in volumes organized chronologically. Lawyers would use the index to find cases relevant to their current dispute. This took time—lots of time. Even with the advent of computer databases such as Lexis Nexis and Westlaw, researching case law was still laborious because you had to try a lot of different word combinations to make sure you weren’t missing a case where a judge used slightly different terminology. Or your search term might be very common and you’d have to read through a lot of cases to find the ones that were most similar to your dispute.
CARA makes this process exponentially faster; she “reads” your brief so she understands the context of your dispute, and then she instantly searches a database of millions of cases and tells you which ones are relevant to your dispute—but she’ll omit the cases you cited in the brief, since you clearly know about those already.
As amazing as CARA is, however, the truth is that doing the case-law research is only part of the battle. If you’re involved in a court dispute, someone still needs to write the brief and show up in court to summarize the brief orally for the judge (among other tasks). There are companies out there, such as ROSS Intelligence, that are testing AI-assisted brief writing, but we’re still a ways off from robots showing up in court and talking to the judge.
In sum, lawyers perform a variety of complicated tasks. Computers can already do some of these tasks much better than humans—but not all of the tasks. Until that happens—or until we make lawyers’ jobs less complicated (perhaps an even more challenging task given the power of inertia)—we will still need human lawyers to wield these impressive AI-powered tools.
This article is part of our bots landscape series. You can download a high resolution version of the landscape featuring 197 companies here.