Towards the conclusion of many legal investigations, lawyers typically find themselves mired in paperwork, documents that span several years if not decades of information pertaining to a specific issue.  Legal teams are then tasked with the seemingly insurmountable challenge of sifting through millions of these spreadsheets, emails, and other documents and sorting them into the necessary categories.  Not only is this process extremely inconvenient and inefficient, it is also incredibly costly and runs the risk of missing important information that is relevant to a case, information that could show that fraud has taken place.

A few weeks ago, SIIA published an AI Spotlight on the benefits of FICO’s AML Threatscore tool which uses artificial intelligence to detect for irregular financial activities like money laundering, terrorist financing, and fraud.  In that same vein, this week’s Spotlight is on AI technology that intuitively aids in the sifting and sorting process of documents for many organizations, including legal organizations, and can help detect for corruption.  Specifically, the company that helps do this is called RAVN.

RAVN (pronounced Raven) is a start-up company based in London that uses various AI platforms to do the sort and sift process that people used to have to do manually.  It does this at a faster, cheaper, and more efficient rate than humans could.  At face value, the ability to do this doesn’t seem like a novel concept since databases are increasingly digitized for ease of use.  However, not only does RAVN do the sorting and the sifting, but it also has the intuition to help guide investigators to data that it thinks is relevant to a case, data that legal teams may not even know is important at the time.  RAVN does this by analyzing for patterns and irregularities in large quantities of data.  Additionally, RAVN has the capability of deciphering formats that may not be uniform, like scans where one of the pages was accidentally scanned upside down, and interpreting them normally.

One example of RAVN’s success is when in January 2016, the U.K.’s Serious Fraud Office (SFO) in London partnered with RAVN in its investigation of Rolls Royce for corruption.  RAVN was able to sift through over 30 million documents at a rate of 600,000 documents a day.  This compares to the human lawyer sifting rate of 3,000 documents a day.  Not only was RAVN faster at doing this, but it was also cheaper and produced fewer errors than its human counterparts.  RAVN’s intuition enabled the SFO to find information that led to Rolls-Royce admitting to bribery and having to pay a £671 million fine.  

Notably, RAVN did not displace the lawyers that had to do the sifting.  Rather, RAVN supplemented the lawyers’ work and enabled them to wade through extreme quantities of documents more cheaply and efficiently, with less risk of human error.  It then presented patterns using that data that the lawyers may not have been able to find or even think to find on their own, but the lawyers had to interact with RAVN.

This is a concept that SIIA highlighted in AI Spotlights past and in SIIA’s recent issue brief titled, “Artificial Intelligence and the Future of Work.”  In many cases, the implementation of AI tools like RAVN, are supplements to human work, not replacements.  This is true in a breadth of sectors, not just in the legal sphere.  RAVN may foreseeably displace some low-skilled workers who perform menial tasks like sifting and sorting documents, but more high-skilled opportunities in the applications of this type of AI are also presenting itself.  It’s important to take stock of these opportunities for high-skilled workers in AI as the world continues to innovate and find creative, cost-cutting solutions to problems that plague society.

Diane Pinto is the Public Policy Coordinator at SIIA. Follow the Policy team on Twitter @SIIAPolicy.