By: Michael Feldman

If ROSS was an actual human, he would certainly be the highest paid legal assistant in history. He has an encyclopedic knowledge of legal texts, and devours new ones quickly and effortlessly. Better yet, he can work seven days a week, 24 hours a day, and requires no office supplies, health insurance or retirement programs. Yes, ROSS is a computer program and his designation as a male is just one of convenience.

The technology behind it was derived from research performed at the University of Toronto, which spun out into ROSS Intelligence by co-founders Jimoh Ovbiagele, Andrew Arruda and Pargles Dall’Oglio. The company has since attracted seed funding from venture capital icon Y Combinator and has relocated to Silicon Valley in California.

ROSS is not just a computer program. The software is an attempt to bring artificial intelligence (AI) into the legal world, acting as an automated assistant, with the intent to augment the work of actual lawyers. It can mine text-based legal documentation and find the appropriate passages and citations than correspond to the question being asked of it. The idea is to enable human attorneys to do their jobs faster and more accurately. According to Arruda, “our aim is to have ROSS on the legal team of every lawyer in the world.”

It joins a growing number of analytics and search platforms that can be used as on online service. Legal research, in particular, offers a rich market, with an estimated value of over $8 billion annually..With that as motivation, a myriad array of services have become available for such work, including LexisNexis, Bloomberg Law, HeinOnline, Westlaw, and JustCite, among others.

Not all, however, are as linguistically advanced as ROSS. Queries may be posed to it in normal English, and if the topic is within its legal scope, it should return an answer within seconds. It is able to cite sources, suggest readings from related legislation and case law, and just generally provide relevant background material for the topic at hand.

If this sounds like the IBM Watson supercomputing platform, that’s because it is -- or rather an instance of it that taps into Watson’s cognitive computing smarts via programming interfaces that IBM makes available to third-party software. The result is an advanced legal analytics platform that leverages Watson’s talent for natural language, information retrieval, and automated reasoning.

According to ROSS Intelligence CTO Jimoh Ovbiagele, the application is comprised of about 100 thousand lines of code, which took approximately 11 months to write. It is not, however, a shrink-wrapped software package that you can purchase at a retail outlet and load up on your in-house desktop or server. Rather, ROSS is offered via a web-based portal that runs on Amazon Web Services (AWS). As a consequence, Amazon gets a cut of the action every time a customer uses ROSS, as does IBM.

The exact licensing details are not for public consumption, nor are the exact number and nature of the AWS compute instances used to power the software. What is almost certainly true is that ROSS will work far cheaper and faster than your average carbon-based legal assistant.

ROSS’s first specialty was bankruptcy law. The software underwent a ten-month crash course on the subject, and at the beginning of May the company announced its first paying customer, BakerHostetler, a firm that that intends to employ it as part of its Bankruptcy, Restructuring and Creditors' Rights team. A few weeks later Latham & Watkins and von Briesen & Roper were also revealed to be clients. Other companies may have signed up as well, but if they did, have opted not to disclose their use of the technology.

ROSS’s potential value to the legal profession is enormous. Its ability to review reams of arcane technical documentation, mine the salient data from it, and spit out the relevant information, could transform the industry. It’s not that ROSS actually understands the intent of case law as a lawyer would; it just makes the correlations necessary to provide relevant responses.

For example, when ROSS was asked “How is an insider defined in bankruptcy code?,” the software returned a highlighted passage it discovered in a case it found from 2012 (Re Winslow, Dist. CT. E.D. N.C.): “The legislative history of section 101(31) supports the In re Broumas approach. An insider as defined in the bankruptcy code is intended to encompass those who have a sufficiently close relationship with the debtor that their conduct is made subject to closer scrutiny than those dealing at arm’s length with the debtor…”

Admittedly, the technology is not at the point where it can replace a lawyer, but it can provide some of the functions of a flesh-and-blood paralegal. Above the Law, an online publication that serves lawyers, thinks the technology could upend the legal profession, not just because it can replace the data mining tasks of legal assistants, but also because it will transform how regular lawyers navigate their profession. In an article published in May, Joe Patrice writes:

“The danger ROSS poses to the legal profession is in disrupting the continuing education of lawyers. From the perspective of the partner (and, by extension, the client), a young lawyer billing 15 hours to collect and analyze case law across multiple jurisdictions to forge a comprehensive memo on a discrete question makes little sense when ROSS over there can do it in a matter of seconds. And serving clients is the whole point of this job and if ROSS can help lawyers do that better, it’s going to become a necessary cog in the legal machine. But all that mental drudgery is what transforms that young lawyer into the future top-billing legal eagle.”

Be that as it may, if software like ROSS can improve the productivity of lawyers, and thereby the bottom line of the firms they toil under, this technology will have a bright future. The larger question of how to transition a workforce and a society where white-collar workers are being digitized away, will have to be dealt with at some point in a more comprehensive manner. Perhaps some innovative young entrepreneur can devise a digital assistant to help solve that problem.