Picture this: You take your seat in your first-year Contracts course. The topic of the day is the Statute of Frauds. After outlining a hypothetical and quickly scanning the seating chart, your professor puts you on the spot: “Under these facts, would the Statute of Frauds preclude enforcement of the parties’ agreement?” Under cold call conditions, even the most composed 1L is likely to feel a pang of anxiety, and you are no exception. But instead of stammering through a recitation of the arcane legalese you poured over the night before, you enter a search into an application capable of identifying and synthesizing relevant case law in your jurisdiction and allow the output to guide your response. A few years later, as a practicing attorney, you apply the same technology to a real-world legal question raised by a loyal client, who is grateful for the savings on billable hours you are able to pass along to her.
Far-fetched as it might appear at first blush, this scenario could be closer to reality than you think, thanks to a team of students at the University of Toronto who created what has been dubbed “Siri for legal knowledge.”
As part of an intra-school competition in December 2014, students from the department of computer science and the Faculty of Information developed an electronic paralegal system called ROSS, aimed at assisting lawyers with case research. The application draws upon the super computing power of IBM’s Watson (which famously competed on Jeopardy in 2011) to intelligently analyze volumes of case law, statutory provisions, and administrative regulations.
While familiar legal research platforms such as Westlaw and Lexis have integrated Google-style plain-text search engine functions, ROSS represents a first step towards true artificial intelligence applied in the legal context. ROSS can predict the outcome of court cases, identify and cite to controlling precedent, and suggest helpful readings on points of law at issue at any stage of a case. Recognizing that the law is seldom completely settled in any area, ROSS also provides a percentage estimation of its own accuracy. Research tasks that once consumed many billable hours could conceivably be reduced to mere minutes, saving legal practitioners time, and clients money.
ROSS is poised to inject the traditionally staid legal profession with a dose of artificial intelligence (AI), and Silicon Valley is taking note: The U of T team placed second overall at an IBM-sponsored cognitive computing competition in New York City earlier this year, and just last week ROSS won backing from the Y Combinator, a company that helps fund and nurture tech startups. The potential implications for the legal community if ROSS gets off the ground are myriad. First, as technologies such as ROSS enter the marketplace, lawyers will have to face the somewhat uncomfortable truth that some aspects of the practice of law can be successfully automated, although it is unlikely that AI will be able to duplicate complex legal reasoning. However, ROSS’s capabilities illustrate the rarity of truly unique legal problems; rather, lawyers are often called upon to employ the same proven approaches in slightly different contexts.
Second, firms might leverage ROSS to decrease the complexity of legal work within notoriously complex statutory schemes and within transactional practice groups facilitating mergers and acquisitions of large, publically traded entities. Younger lawyers working in these fields could take on more substantive work earlier in their careers as they acquire expertise not only through experience but also through AI technology. While the integration of ROSS might reduce law firms’ hiring needs, attorneys at the bottom of the firm hierarchy will be less expendable if they embrace this new resource.
Third, ROSS’s impact on law school instruction is potentially consequential, because students will have access to an adaptable, intelligent research tool when working casebook problems or responding to hypotheticals in class. The effect of introducing a powerful facilitator of legal knowledge could parallel the appearance of the calculator in algebra classrooms in the 1970s. Like the advanced math student unencumbered by basic arithmetic, the law student of the future could spend less time digesting cases and more energy applying the facts of a hypo to the relevant law. In short, law students will be trained to think like practicing lawyers sooner.
AI technologies such as ROSS represent an important aspect of the future of the legal profession, from the novice law student to the seasoned partner. Whether in Contracts or in the courtroom, AI promises to change the legal landscape, but as with all innovations, exactly what changes will result remains unknown.