How artificial intelligence is transforming the legal profession
The technology boom was just beginning with the emergence of email and personal computers. Jay Leib was working for Record Technologies Inc. as director of software sales and training in 1999, and the company was scanning documents into databases for clients. At one point the company printed and scanned legal documents related to a lawsuit with Microsoft. Leib thought that was inefficient, a waste of time and paper.
So he and his business partner, Dan Roth, decided to create a program that would help lawyers manage electronic documents for litigation. Their idea led them to purchase an e-discovery application. By 2000, Leib and his partner launched their own creation, Discovery Cracker.
“We saw a gap in the marketplace,” Leib says. “Why print all that paper? Lawyers need tools to keep up with it.”
Instead of wading through piles of paper, lawyers now deal with terabytes of data and hundreds of thousands of documents. E-discovery, legal research and document review are more sophisticated due to the abundance of data. So while working as chief strategy officer at kCura in Chicago, Leib saw a need again in the market.
“What is the future of the industry? We thought about it,” he says. “There were whistleblowers in their companies who knew what was going on, and the unstructured data contained the stories. Companies could detect potential problems early on, provide alternatives to counsel and C-suite, and understand their exposure. It would prevent unnecessary legal spend and mitigate risk, thus protecting the company’s brand and shareholder value.”
In 2013, he and Roth, a professor and head of the cognitive computing group at the University of Illinois at Urbana-Champaign, created NexLP, a company using artificial intelligence to analyze data and identify trends.
A REVOLUTION BEGINS
Artificial intelligence is changing the way lawyers think, the way they do business and the way they interact with clients. Artificial intelligence is more than legal technology. It is the next great hope that will revolutionize the legal profession.
Change can be brought on through pushing existing ideas. What makes artificial intelligence stand out is the potential for a paradigm shift in how legal work is done.
AI, sometimes referred to as cognitive computing, refers to computers learning how to complete tasks traditionally done by humans. The focus is on computers looking for patterns in data, carrying out tests to evaluate the data and finding results. Chicago-based NexLP, which stands for next generation language processing, is creating new ways for lawyers to look at data.
“Gerry Spence once said, ‘Telling a story is one of the most persuasive means of communication,’ ” Leib says. “Text analytics and machine learning can be incredibly helpful in helping the data tell its story, thus allowing legal teams and the C-suite to focus their time on nuanced analysis and application of that story to the issue at hand.”
Leib is not interested in the usual data analytics but, rather, in preventive measures, including predicting litigation and measuring workflows in real time. His company uses predictive coding, whereby users sample data and identify what is relevant. Through sampling, the program is able to learn what documents are relevant. This process greatly reduces the time needed for e-discovery and document review because the program is searching for concepts as opposed to simple keywords. The company is interested in identifying key information to predict future outcomes.
“Analytics can help in many areas of a business, not just legal,” says Leib. “We can guide compliance departments to streamline internal investigations to get to key information within hours.
“IT professionals have also been pressed into investigations of data breaches,” he notes. “In the Sony data breach, unstructured data was exposed that was financially damaging and embarrassing, underscoring the need to be in front of it to understand what insiders are discussing within the four walls of the corporation.”
Storytelling is one of the features of NexLP’s work.
Deep within the data lies a story, whether it’s a story to tell a judge at trial or to pitch to potential clients.
And there is an enormous amount of data that is being generated. According to IBM, 2.5 quintillion (2,500,000,000,000,000,000) bytes of data are created every day, and 90 percent of all data was created within the last two years. In order to tell a good story, lawyers need a way to sift through the data.
“Nearly 80 percent of a company’s data is unstructured,” Leib says. “While unstructured data represents the lion’s share of a company’s data, for years lawyers have been stuck with antiquated tools that focus primarily or solely on Boolean search. Better tools are needed to truly understand data, infer meaning, classify the various types of ideas present, and help you get to the result fast—even if that result didn’t involve the keywords you used.”
Roth helped develop technology that can turn information into stories. Story Engine is a program that can read through unstructured data and summarize conversations, including the ideas discussed, the frequency of the communication and the mood of the speakers. The company uses the data to build models to analyze behavior and find signs of fraud or litigation.
“For example, when investigating securities fraud, price movement can be a very useful indicator,” Leib says. “However, stock prices rise and fall throughout the trading day. Our analytic engine can overlay communications between traders discussing that stock-on-top-of-price-movement data to compare the times they both occurred. Perhaps the traders in question also emailed client information to themselves. By com-paring these various data points, a clear pattern can quickly emerge—one that might have previously gone unseen or would have been considered circumstantial. These patterns allow financial firms to better understand and identify this behavior to prevent it, and also tell easy-to-follow stories to regulators or judges.”
SEEING THE FUTURE
Another potential use for data is predicting legal outcomes. In 2014, Chicago-Kent College of Law professor Daniel Martin Katz, then at Michigan State University law school, and his colleagues created an algorithm to predict the outcomes of U.S. Supreme Court cases. It attained 70 percent accuracy for 7,700 rulings from 1953 to 2013. Leib’s company wants to take this idea one step further, working with analyzed information to predict future litigation.
“As companies develop better metrics around things like litigation and compliance spend, the barriers to entry that analytics tools used to face are quickly falling by the wayside,” says Leib. “Once companies get their arms around case flow and spend, the ROI and economics around using analytics to streamline legal workflows and reduce or eliminate risk becomes much more attractive.”
NexLP offers services where clients use the software to identify patterns in the data. Once any possible issues are flagged, the system can collect the necessary documents for any possible litigation.
“The biggest differentiator for us is our strength in pattern recognition,” Leib says. “We will be able to detect problems ‘in vitro,’ … closer to where they’re happening. We surface and expose anomalies people should be paying attention to. We can take key issues and fact patterns common to certain types of matters and build models that identify and prioritize documents that should be looked at first. For example, for trade secret theft, we can identify behaviors that can quickly pinpoint the time frame the theft occurred, how it was accomplished and who was involved.”
Leib wants lawyers to think differently about legal technology. His company is working on measuring emotional responses and using existing technology such as programs designed to detect insider trading.
“More and more every day, we are building the bridge to increased analytics usage,” says Leib. “Our clients have grown comfortable using keywords. I believe that keywords are only one part of the equation, and keyword usage alone leads to inefficiency, increased cost and unnecessary risk.
“AI classifies and organizes data faster, better and cheaper, and augments human intelligence,” he says. “It empowers people to make use of huge amounts of data to make better decisions and tell better stories.”
Legal experts such as Richard Susskind and Jordan Furlong have long been writing about the legal profession’s woes and its stubborn adherence to traditional culture. The Canadian Bar Association laid out the need for legal reform in a 2014 report, Futures: Transforming the Delivery of Legal Services in Canada (PDF). The report stated that the key to a viable, competitive and relevant legal profession is innovation.
So what the profession could use is an industry leader willing to take a calculated risk. This is where Dentons steps in.
In 2015, the world’s largest law firm created NextLaw Labs, an independent subsidiary designed to disrupt the legal industry through innovation. So far their plans are working: The Financial Times recognized Dentons as the most forward-thinking North American law firm last year.
Dan Jansen, an entrepreneur with a background in advertising and media, came on board as the first CEO of NextLaw Labs. After years of making lawyer jokes to his wife, who worked as a corporate attorney, Jansen now has the job of transforming the legal profession.
“What drew me is the opportunity of reinvention,” says Jansen. “Law firms have historically had a pyramid structure that technology is evolving into a diamond. If the work at the bottom of the pyramid is being automated, we want to own that technology and not be a victim of it. If this was the advertising industry, we would be 10-20 years too late, but we think we’re right on trend in legal tech.”
NextLaw Labs stands out among the world’s large innovative firms. Even the structure of the company is unique: As an autonomous subsidiary of Dentons, it has the freedom to operate outside of the partnership model, which can be an obstacle for innovation.
“It would be hard to do this internally in any business,” Jansen says. “The last business I had was an alternative advertising model, and it was a similar autonomous entity. We’re sponsored by the firm, and they understand that there are long sales cycles in law and we need to speed that up. We have separate management and separate governance so we can move quickly and drive change quickly in this $600 billion [global] industry.”
According to the 2016 Report on the State of the Legal Market (PDF), published by Georgetown University’s Center for the Study of the Legal Profession and Thomson Reuters’ Peer Monitor, demand for legal services was “essentially flat for 2015 … [and] continues a pattern seen over the last six years.” Also “there has been an overall downward trend in the productivity of all categories of timekeepers except associate.” The report attributes at least some of this stagnation to business clients’ reduced spending—a jaw-dropping 25.8 percent between 2004 and 2014 in inflation-adjusted dollars, the 2015 report noted. With increasing competition in the legal market, law firms are under pressure to invest in innovation.
U.S. businesses generally spend, according to Jansen, about 3.5 percent of revenue on research and development. “The legal industry spends less than 1 percent on R&D, compared to telecommunications for example, which spends 13 percent, or biotech, which spends even more,” says Jansen. “We can have a real impact in the industry with modest spending given where the legal tech sector is at. It’s an opportunity that is fairly unique.”
What makes NextLaw Labs special is its drive for change. Jansen and his staff began by consulting with lawyers and staff at Dentons for ideas.
“So far, 80 percent of the ideas we spend our time on are from our partners or clients,” Jansen says. “Actually, we have partners who say: Here’s a problem, here’s a possible solution and here’s the prototype that I’ve been thinking about. There are joint ventures involved, business launches, product launches, and we can help accelerate that while contributing to new and different ways that lawyers can work with their clients.”
The NextLaw Labs advisory group weeds through the top ideas to pick out those that are most promising, and it then conducts market research to see whether others are working on the same ideas. If they find a large company is involved, Jansen says, they seek to partner with it, and if it’s a small firm, they’ll provide investment capital and advice.
“We also have access to a cognitive computing platform with IBM Watson that provides a menu of tools that represent very sophisticated technologies, which is also attractive,” Jansen adds.
The abilities of the Watson question-answering supercomputer have also drawn the interest of legal industry giant Thomson Reuters. Eric Laughlin, a managing director at Thomson Reuters Legal, announced in late January that the Watson Initiative he heads hopes to have a beta version of a corporate compliance product available for testing later this year. And Robert Schukai, head of applied innovation, technology and operations for Thomson Reuters, says the firm is building a center for cognitive computing.
Meanwwhile, one NextLaw Labs startup that immediately stands out is Ross Intelligence. A Canadian partner in Dentons (along with IBM) informed Jansen about the Toronto-based firm. Ross Intelligence uses the Watson cognitive computing system to enhance legal research. Users ask legal questions in plain English and Ross searches legislation, case law and secondary sources. NextLaw Labs signed Ross last August as its first portfolio company, providing capital and office space in Palo Alto, California.
“We resolved to find and work with innovative law firms who saw the future and see something big,” says Andrew Arruda, a co-founder of Ross Intelligence. “With NextLaw Labs, their financing was just one of many factors that showed their commitment to innovation to the law. It’s great to be part of the initiative.”
Ross began as an idea from co-founder Jimoh Ovbiagele, who was deeply affected watching his parents struggle to pay hefty legal fees due to expensive legal research. Like NextLaw Labs, the goal is more than creating good technology; the firm is working to make legal research easier and more affordable and thus bring down legal fees for consumers. In its short time with NextLaw Labs, Ross has gained 20 clients in the United States and has plans to expand to international markets.
“Ross Intelligence is not to replace lawyers but to [allow them to] do more than they were able to do before,” Arruda says. “We’re working on having lawyers teach the computer to think like a lawyer. That would be a huge step for humanity.
“With legal tech, there will be new jobs, and we can embrace a very happy future in the law,” he says. “This is a new frontier.”
Most legal startups have difficulty raising capital because of the unknown market and the long turnaround time required for product development and launch. NextLaw Labs partners with venture capital firms to help young companies grow.
“What we found working with venture capitalists is that they appreciate our expertise and Dentons as a potential first major customer who can help to shape their nascent companies’ offerings,” Jansen says. “Our lawyers appreciate that we’re funding these companies that can give them access to cutting-edge solutions that they can introduce to their clients.”
THE ARTIFICIAL INTELLIGENCE GENERATION
Adam Nguyen suffered from a career problem that many lawyers have faced. The Harvard Law grad began his career as an associate at Shearman & Sterling in 2002, working late nights doing due diligence and copying and pasting information into various legal documents. He switched into many different roles, working as a clerk at the U.S. District Court for the District of New Jersey and later as an in-house lawyer at AQR Capital Management.
“I thought about what I should do with my life,” Nguyen says. “As a group, lawyers tend to be miserable because a significant portion of our work is tedious and doesn’t require much analytical skills. Many of us hadn’t entered the profession expecting to spend much time on tasks like copying and pasting, document summarizing or assembling document binders.
“Moreover, clients increasingly are pushing back on paying for less sophisticated, manual tasks,” he notes. “Technology that automates tedious tasks, while not a panacea, can free up lawyers’ time to perform higher-level, more intellectually satisfying work which clients would be willing to pay for. It would help to make lawyers happier and more productive. There are many smart, eager young lawyers whom we’d want to keep in our profession.”
Nguyen wanted something better. So he and co-founder Ned Gannon approached Columbia University to work on technology for due diligence and contract review. The technology already existed, but it hadn’t been used much in the legal field. They worked on the technology for nine months, and eBrevia’s Diligence Accelerator program was launched in 2012.
Companies like eBrevia aren’t seeking to eliminate lawyers, but to make their lives better. Law firms are feeling the pressure from clients, particularly in-house counsel, to lower costs. And artificial intelligence is born out of necessity.
“Companies and their lawyers often have to perform a cost-benefit analysis in areas like legal due diligence,” Nguyen says. “Perform a partial review of a database of documents and keep costs low, or perform a thorough review and blow through the budget? It’s clear we need a solution that increases lawyers’ productivity while helping to deliver legal services at the level expected by clients.”
The focus for eBrevia’s products is on extracting information from data. Clients upload documents to the server, where they search for information and download it sorted according to their preferences. What makes eBrevia different is the program’s ability to learn how to do searches more efficiently.
“The machine-learning technology learns from examples,” says Nguyen. “It’s been taught to recognize and abstract legal concepts like assignment, change of control or renewal. The user would upload documents to the system, and it would return sentences relating to legal concepts that it has learned.
“Many users ask: Isn’t this just a word search? The answer is no. Among other things, AI learns from the relationships between words and what provisions and concepts look like from samples of documents. After learning, the software doesn’t have to be retrained for a new transaction as with litigation e-discovery tools. The software recognizes legal concepts regardless of the vocabulary used to express them. And it’s adaptable, because adjusting to a new legal domain simply requires training on that domain’s examples.”
The company hopes to expand to new markets. It partnered with RR Donnelley and introduced Venue contract analytics in October. The program allows RR Donnelley’s Venue virtual-data-room users to have automated contract review and due diligence using eBrevia’s artificial intelligence software. The company also plans to continue expanding its reach by developing products for real estate and M&A deals.
“I was on a law panel in March 2015 and the title was something like ‘Will robots take our jobs?’ ” Nguyen says. “No, it’s not about replacing but transforming. About 20 to 30 years ago, lawyers had to manually red-line documents, but we don’t do that anymore. We don’t perform legal research by going to the books. Do we still have to red-line documents and perform legal research? Yes, and actually we are doing a lot more. I would argue that technology transformed our jobs and enabled us to … add more value.”
While Nguyen has his sights set on Europe, U.K. firms have their eyes set on North America. The AI movement is more advanced in the U.K., with firms partnering with universities and using financial incentives to innovate. Alternative business structures have been around since 2007 there, opening the door to new competition for law firms.
Again, pressure is on large firms to cut legal costs and become more efficient with less time and money to work with. The result is traditional law firms, accounting firms, legal startups and myriad other businesses are competing for clients.
Ravn Systems is in the next line of U.K. companies looking stateside. Founded in 2010, it focuses on using cognitive computing to extract information from data with its Applied Cognitive Engine program.
“We derive structure from chaos,” says Ravn CEO and co-founder Peter Wallqvist. “For example, we can tell what are the tables, price lists, clauses and charts. We look at that with the robot and transfer it into structured data. What we do very well is having the ability to understand every little bit of the contract.” While Ravn’s work is similar to other companies in the AI field, its vision reaches far beyond the law. The company’s next target is the global financial markets.
“The OTC trading market is at least 10 times bigger than exchange-traded security markets,” Wallqvist says. “Selling contracts to others is unbelievably painful for banks. Had people used this kind of technology in 2006 or 2007, maybe they could have known how unsustainable the market was. That was the OTC trading market and the banks didn’t know how much exposure they had. There’s a huge opportunity to extract the data from the structure and provide more information.”
Artificial intelligence has the potential to bring in new business for lawyers, too. Because Ravn’s technology can quickly go through large amounts of data, it opens the door for different legal work that previously wasn’t possible.
“For example, there’s a firm with a corporate client,” says Wallqvist. “There is a new law in the EU that says an employer is liable for overtime pay even during the employee’s paid vacation days. When you’re a company with 50,000 employees, that’s a big deal. The client asks the firm’s employment lawyers: How much will we be liable? The firm says: We would have to read through all the documents and it would take one year for two lawyers to do that. Now they can instead use our technology and do work like this,” he says, predicting his firm could do the job in a few days. This legal revolution is moving quickly. Ravn announced in December that its new Ravn ACE Contract Robot would be able to extract data from title deeds along with other documents.
Riverview Law, a U.K.-based firm known for its progressive ideas, recently launched its virtual assistant Kim, which stands for knowledge, intelligence and meaning. The program will use artificial intelligence technology from the University of Liverpool and Clixlex, a U.S.-based data collection and management program that was acquired by Riverview last summer and renamed Kim Technologies.
While AI is growing, it hasn’t reached the majority of law firms yet. What will really make artificial intelligence a revolution is to change the thinking of lawyers. Perhaps real change will come with a simple recognition of the need to better serve clients.
“You start with a number of documents and ask questions like what are the termination clauses,” Wallqvist says. “For example, there’s a major telecommunications company that would tell us about documents they had to review. They told us how they had to go through 1,000 documents, which would take three people six months to complete. We can do that in a matter of days.”
That is the future. Maybe it’s not so scary after all.
This article originally appeared in the April 2016 issue of the ABA Journal with this headline: “Beyond Imagination: How artificial intelligence is transforming the legal profession.”
What Is AI?
Artificial intelligence is the legal tech buzzword of 2016, but it can be misunderstood. AI, also called cognitive computing in the legal tech world, is about machines thinking like humans and performing human tasks.
“Cognitive computing enables robots to learn,” says Garry Mathiason, co-chair of the robotics, AI and automation industry group at Littler Mendelson in San Francisco. “In traditional software, the possibilities are mapped out and predetermined. This has limited the development and application of software-driven machines and robotics.
“However, this is dramatically changing with the introduction of cognitive computing. Modeled after human learning, smart machines process massive data, identifying patterns. These patterns are used to ‘create’ entirely new patterns, allowing machines to test hypotheses and find solutions unknown to the original programmers.”
There are two types of artificial intelligence—hard and soft. Hard AI is focused on having machines think like humans, while soft AI is focused on machines being able to do work that traditionally could only be completed by humans. The main difference is that soft AI doesn’t necessarily involve machines thinking like humans.
“Our perspective on artificial intelligence has changed significantly over the past several decades,” says Jack Conrad, lead research scientist in corporate R&D at Thomson Reuters and president of the International Association for Artificial Intelligence and Law. “AI failed to live up to the early expectations that focused largely on hard AI capabilities, such as the ability to perform humanlike reasoning. When those lofty goals were not attained, researchers came to understand how difficult such achievements really were. After all, trying to teach computers to perform cognitive activities was an extremely challenging task.
“Over time, as expectations were lowered and research efforts became more narrowly directed, a shift towards ‘soft’ AI applications took place, focusing on providing intelligent tools and problem-solving resources to humans.”
Laws and Robotics
Many legal and ethical issues surround the use of AI. For example, the issue of liability hasn’t been resolved yet for Google’s driverless cars, though researchers such as UCLA professor John Villasenor and others argue that product liability could cover any driverless car accidents. And in employment law, when a machine used to interview potential job candidates measures the person’s blood pressure, privacy becomes an issue.
“When AI is coupled with big data, the solutions formed can unintentionally conflict with workplace laws, some of which were written 50 to 100 years ago,” says Garry Mathiason of Littler Mendelson. “For example, big data shows that the closer one lives to where one works correlates with job longevity. When this finding is discovered by a brilliant machine evaluating applicants, it is likely that those hired will disproportionally live close to where the company is located. What happens if the company is located in a rich, nondiverse neighborhood? This seemingly benign process may be having a disparate impact on the diversity of those hired, leading to a claim of racial discrimination.”
The current regulatory scheme doesn’t account for machines, but researchers, lawmakers and academics are working on the different issues. Mathiason believes bar associations can help.
“There’s no way for workplace laws, regulations and court decisions to keep up with the technological changes,” Mathiason says. “Members of the bar are called upon to provide soft-law advice and solutions that will bridge the gap between today’s workplace decisions and tomorrow’s review of those decisions by courts, regulators and legislators.”
Julie Sobowale, a lawyer, is a freelance writer based in Halifax, Nova Scotia.