Bring up artificial intelligence (AI) in a room full of lawyers and you’ll typically get two reactions. The first is a scoff, usually accompanied by sounds about how AI will never replace lawyers because the law is about relationships, it’s about human interactions and needs human judgment. For many years, as other industries such as manufacturing and finance have experienced process and job disruption attributed directly to AI, yet the legal world has not, this has held true.
However, look carefully and you’ll find the second type of reaction, typically coming from the younger generation of lawyers, the ones who have graduated more recently. These are the ones who are less experienced in the law but more experienced, perhaps, with technology and its implications for the future. From this crowd, you are likely to see some crinkled brows and nervous laughter. They know that Rosie from the Jetsons isn’t about to stroll into their offices tomorrow and supplant them, and yet… They do know that machine learning is making inroads even in the legal industry.
The good news is, Rosie isn’t about to get her juris doctorate anytime soon, and probably not within any of our lifetimes. But task-by-task, AI and the companies that develop it are taking aim at the legal field, where there is ample space for improving efficiency and finding cost-savings. After all, efficient and cheap aren’t two words commonly associated with legal work. Many startups working with AI believe that should change, and are working diligently on making that change a reality.
Working Alongside AI
In January 2017, the McKinsey Global Institute released a report, A Future That Works: Automation, Employment, and Productivity.
Among other things, a major conclusion of their report is that in most industries people will need to work alongside machines in order to ensure productivity. And the fields most susceptible to AI and to automation, in general, continue to be those that rely on physical activities – notably, legal work does not fall into this category. However, just because the legal field is not the most susceptible market to AI disruption does not mean that it is impervious to it. In fact, for those who have been paying close attention, the disruption has already begun.
Unbundling Legal Work in Order to Automate It
Technology such as AI is slowly unbundling legal work into distinct tasks and identifying those that are susceptible to being performed with no or minimal human input. Most legal experts anticipate that this process will take at least a decade or two. The most highly skilled lawyers, working on bet-the-company litigation or complex deals, are fairly safe from job disruption by AI.
However, baseline legal services, such as many of those currently performed by paralegals, legal assistants, and junior associates, will continue to be slowly broken down into tasks that can then be partially or fully automated with the help of AI programs. As AI gets better at continuously learning and self-correcting, the process of unbundling is likely to accelerate exponentially.
Which Tasks Are Currently Unbundled?
Any talk about “unbundling” legal tasks begs the question of which tasks are currently undergoing a disruption by AI. For the most part, these are tasks that are of the search-and-find variety, including e-discovery and legal research. There is no shortage of companies, including dozens of startups in the last two years alone, willing to spend time and money on automating these sorts of legal services.
Companies such as DISCO and Brainspace are helping legal clients find documents and relevant metadata faster and more accurately than ever before. On the one hand, applying AI to e-discovery problems has been an effective cost-saver for both law firms and their clients. On the other hand, however, there is no denying that without AI these are tasks that would allow more hours to junior lawyers and legal paraprofessionals.
It’s also worth noting that the rise of technology as evidence, as well as emailing, text messaging, and cloud computing as forms of client communication, has resulted in volumes of e-discovery that can be virtually impossible for humans to sift through, even given unlimited time to do it in – not that clients are willing to pay for unlimited time. The best AI disruptors in the e-discovery field are those who have not forgotten about the human touch and use it hand-in-hand with AI machine learning. When the first review is performed by AI and the second by skilled humans, clients benefit from both efficiency and necessary judgment calls.
Just as the sheer size of e-discovery in the average case is increasing, as time goes on the case law relevant to any given case is also growing. Thousands of cases are added to databases every day just chronicling U.S. opinions in state and federal courts.
Through the implementation of AI, searches that might have taken hours can be performed in mere minutes, with connections brought to light that would be difficult, if not impossible, for a single lawyer or team of lawyers to make. For example, Knomos, a Canadian legal research, and technology company, uses proprietary software to analyze case law research results and then map them out into a 3D visual representation that leads lawyers to make significant connections between obscure data points.
When you apply machine learning and predictive analytics to legal content, you not only see patterns within the texts themselves but can also trace patterns in the metadata; to see which jurists and scholars are being cited most often, which arguments prevail in different types of cases, and the ebb and flow of leading ideas as the law evolves – Adam La France, CEO of Knomos Knowledge.
The implementation of AI appears almost unlimited, even if today’s firms are using it only in a few areas. Some startups are already coming up with unique ways to use AI. According to Intraspexion‘s CEO, Nick Brestoff, his company is all about prevention. Using a form of AI called Deep Learning, Intraspexion extracts text from specific types of lawsuits (such as breach of contract, fraud or discrimination), builds a model for each pattern, and then looks at and flags the risky emails for review by corporate counsel.
Our patented system provides an early warning, Brestoff told us. That's a high ROI, to be able to find the risky or smoking gun emails before a lawsuit is filed, rather than to have to manage the lawsuit after the damage is done.
But if there is a lawsuit, Premonition has built the world's largest litigation database and uses it to reveal which lawyers win in front of which judges. This type of data is tremendously valuable to a number of industries, especially the insurance industry as it leads to better risk assessment and claims management
Our goal is to bring transparency to the previously opaque industry of law. This will result in greater efficiencies and better outcomes. Every other industry relies on metrics and data to improve performance and decision-making. It’s time the legal industry did the same – Nathan Huber, Director at Premonition.
As AI use increases across all fields and industries, so do the legal and ethical issues surrounding its use. Intriguingly, the lawyers who litigate these issues in the future may do so with the help of AI in the process.