Artificial intelligence (A.I.), often referred to in terms like “computer learning” or “deep learning”, is nothing incredibly new. Science fiction authors have dreamt and wrote about the idea of artificially sentient machines or robots since long before the age of modern computers. It’s a fascinating concept, and one that continues to inspire writers, engineers, and technology experts every day. Can something we create gain the ability to truly learn on its own?
As fantastically odd and nutty as many of our favorite fictional A.I. creations were and continue to be, they weren’t too far from the 100% real artificial intelligence of today. We may not have clunky robots making and serving our coffee (come on, give Starbucks a little more time), but we do have computer technology that is partially self-learning and able to make future predictions based off previous knowledge bases. Essentially, we’re beginning to master the fundamentals of artificial intelligence.
Computer learning technology helps us with nearly every task we complete in 2016. Self-driving cars continue to make headlines on a daily basis as they make their way onto our streets. Computer-guided safety systems, vehicles, and transportation systems have saved lives every day. Nearly everything we do on our personal computers, smartphones, and tablets is at least partially guided by A.I. and deep learning processes, in fact.
Deep Learning: As Common as Our Favorite Apps and Websites
Computer learning or deep learning can be found everywhere we look online. Google uses computer learning to decipher our search terms and figure out what we really want to search for at 2 am on a drunken Saturday night. How does Google know how to decipher typos? By learning from information derived from other searches. This is computer learning in one of its simplest forms. This may seem quite different from the concepts of artificial intelligence we’ve come to know from science fiction, but it counts.
Essentially every time a program or software knows how to “predict” human behavior based off previous actions and/or data, that program or software “learns” how to make better predictions. Google’s servers had no knowledge of Pokémon Go until the game swept the world by storm, for example. After one mere day of seeing search engines everywhere explode with Pokémon this and Pokémon that, search engines quickly learned what all the fuss was about and knew how to direct folks to the links they wanted to see.
We can find computer learning-guided behavior on YouTube, on Amazon, on Spotify, or any of our favorite retail sites, media sources, or news sites. Every time we take a look at a list of recommended stories, articles, songs, apps, or products, we’re looking at a list that was generated using deep learning. “What did those other people like?” the program asks itself. “Maybe this person will like this too—especially since they also liked Y and Z.”
When coded and created well, computer learning and deep learning become invisible. This is one reason why we tend to not think of platforms like Google and Spotify using artificial intelligence. When we make predications and judgement calls as human beings, the process is completely natural, after all. When a computer makes those same predictions and judgement calls in an entirely natural, human-like manner, we have a hard time realizing those behaviors are originating from something non-human.
As common as computer learning is becoming, it should come as no surprise that all of our professional industries embrace it in various forms. Computer/deep learning is even regularly used in the legal profession—in more ways than you may think. When a law firm uses a case or document management app, they’re potentially utilizing concepts of artificial intelligence and computer learning. When we talk about e-discovery, we’re talking about discovery methods which utilize computer learning.
Curious to know more? Here are six ways artificial intelligence is helping our lawyers—without them even noticing:
One of the most integral tasks to completing cases is information retrieval, often referred to as information management within the legal technology world. Information retrieval includes tasks like sorting through thousands of business emails to find incriminating evidence or searching through hundreds of websites and social media profiles to find possible witness connections.
In the past, mundane tasks like this had to be handled manually over multiple days or weeks. With the right type of technology and a bit of A.I., however, hunting for the exact data your case needs can become a simple task—one that can be performed while your lawyers are off doing more important tasks. Information retrieval is naturally part of the e-discovery process, but it can encompass other, broader tasks as well.
We had the pleasure of interviewing Peter Wallqvist, CEO of RAVN Systems, about RAVN’s A.I.-based platform. The platform is called RAVN ACE (Applied Cognitive Engine), and it powers multiple applications that legal professionals use on a daily basis to automatically organize, discover, and summarize information retrieved from legal documents.
RAVN ACE is so advanced that it’s been used to power a top international firm’s “Contract Robot” that’s used to extract real estate practice data, check it against external sources, and structure it into a manageable format. The robot takes manual, mundane tasks that would have taken human teams more than 100 days to complete and completes them in less than two seconds. If that’s not an example of work efficiency, we’re not sure what is.
When asked about how lawyers specifically benefit from RAVN ACE, Peter Wallqvist had this to say:
In the case of law firms, junior lawyers can now focus on more interesting/billable work instead arduous manual tasks. They can also mitigate risk by eradicating typos and human errors. In addition, as different people interpret data in inconsistent ways RAVN ACE can ensure that all data extracted is consistent.
With the implementation of RAVN ACE, organizations have seen the added benefit of increasing staff morale, since employees no longer have to perform uninteresting, repetitive data extraction tasks. Instead they are able to focus on more engaging work.
Some of the tasks ACE is able to take over are so dull that it is hard to even find staff willing to do them these days. So in some cases for our clients the choice is either to use technology or simply not accepting the business at all.
Documents are an essential part of any legal practice, and in order to be successful at managing a legal firm’s documents, that firm must have a process by which to organize documentation, easily find information from crucial documents, and create new documents with ease. Using a broad definition, this all fits into a task legal professionals refer to as document management.
Smaller legal practices may be able to keep up with document management on a manual level, but for larger firms or corporate practices, A.I. can be used to analyze documentation for critical information, find errors, or run statistical checks to help that firm complete cases with more efficiency. A.I. can also be used for processing forms like new client contracts, making client profiling and upkeep extremely simple.
We spoke with Ned Gannon, CEO of eBrevia, a technology vendor that specializes in artificial intelligence-based contract analysis, lease abstraction, and document diligence acceleration. We wondered, specifically, how A.I.-based document management improved upon traditional forms of document management. This is what he had to say about eBrevia in particular:
eBrevia uses machine learning (ML) and natural language processing (NLP) technology to recognize complex concepts, like “change of control" or “assignment,” within a contract. Instead of using search terms or having an expert write boolean rules, ML software learns from human-annotated example documents and builds statistical models of each concept.
Because these statistical models take into account a diverse set of features, our software is much more robust than one using keywords or heuristic rules. The software can identify concepts even if expressed in novel language, or if important words are redacted or illegible. The ML technology extracts concepts regardless of the words used to express them, while leaving behind uses of “keywords” that are not actually on point.
Our software can also easily discriminate between different senses or meanings of the same word or phrase. In addition, as the models driving the software are statistical, we can predict the probability that a passage is relevant, and tune our system threshold to manage risk based on the provision type.
Under the right conditions, A.I.-based legal platforms can save law firms and corporate practices dozens of hours each and every week—all the while making sure no errors slip through the cracks.
Another central aspect of all legal practices is the task of employee management. How do you manage issues like employee compliance in a corporate setting without seeking advice from your company’s internal legal department? What about issues of overtime, etc.?
A.I.-based employee management platforms can reduce the need to have an internal team for compliance issues altogether. All legal issues concerning information governance can additionally be further simplified since the program will be able to predict behaviors and know how to react in certain situations. One of the many strengths of A.I.-based platforms is the fact that most platforms are created to predict and react accordingly, helping a law firm prepare for whatever may come their way.
We spoke to Lori Brown, President of ComplianceHR. ComplianceHR has created a platform of applications that contain hybridized reasoning methods. These methods are based off if/then mappings, decision trees, situation sets, multi-value variables, instances and weighted scorings, and more, giving employers the freedom to deal with compliance issues at the core.
Essentially, ComplianceHR applications remove the need to have a team of internal attorneys standing by in regards to compliance. The platform knows how to deal with compliance issues based off knowledge it was granted as well as complex coding that allows it to predict the outcomes of various issues before they arise.
This is what Lori Brown had to say regarding the future of the legal profession and how it will change with the continued advancement of artificial intelligence:
AI is impactful technology, enabling online solutions that provide resolutions using content as input to complex legal reasoning and generating highly accurate assessments and advice on multifaceted matters. Clients gain the benefit of actionable advice without calling an attorney, yet knowing they have guidance based on intelligent experienced-based reasoning.
Adopting a design thinking approach to legal service delivery, as well as using smart technology enable the advancement of solutions like ComplianceHR that create countless benefits in the delivery of legal solutions. As clients continue to demand and expect accurate, cost effective and creative legal advice, the use of AI technology will gain momentum in the legal industry and become mainstream.
E-discovery is one area of the legal profession which sees a large amount of artificial intelligence-based platforms and software. E-discovery is how legal practices are able to locate incriminating evidence among thousands of corporate emails, texts, social media platforms, and web pages. Information retrieval, as mentioned above, is a major part of e-discovery.
One other major aspect of e-discovery is document review. In the document review stage, a team of experts or an A.I.-based review platform goes over all information which was discovered, analyzing the results, and performing new searches based off those analyses. This stage essentially provides quality control. Was the right data located? Is there anything potentially missing?
As you can imagine, an automated platform that performs this crucial step using artificial intelligence can save legal groups a great deal of time and money. Since such platforms perform tasks so quickly, it’s also possible for the platforms to run multiple searches and QC checks, ridding the data of any possible errors.
We spoke with Keith Whitaker, Chief Business Development Officer of Percipient, LLC, about Percipient, a document review platform that uses A.I. to quality check discovery data. He had this to say regarding Percipient’s QC-based engine:
One way we employ AI is for Quality Control during document review. As the review is completed, we’ve built an engine which flags documents that may have been mis-classified. This could be potentially responsive documents marked as non-responsive or documents likely to contain privileged information. Flagged documents are added to the QC set for secondary review. This greatly increases the chance of finding documents that were missed compared to relying exclusively on a random QC set.
At a technical level, the QC engine works by reading a subset of the documents to “learn” the key characteristics in the text and metadata that make documents responsive or privileged. It then applies this “knowledge” to the remaining documents and repeats the process several times. This iterative approach is fully automated and flagged documents remain in the primary review platform. This means no disruption to the workflow, while leveraging the collective knowledge of all the reviewers. The result is a production set that the legal team can have more confidence in.
Law firms both large and small can gain benefits from using case management software that utilizes artificial intelligence and computer learning. What many of these platforms do is help lawyers and paralegals work through the necessary steps of bringing a case to settlement by way of automatic document creation tools and customizable user prompts.
Case Pacer is one such example. Case Pacer uses user prompts called “ticklers” to alert users about important documentation and tasks that need to be obtained or completed and in what order. The software has essentially learned what steps an attorney needs to take in order to bring a case to completion and will predict what actions need to be taken next based on that knowledge.
This is a useful tool for attorneys and paralegals who want to bring cases to settlement with as little hassle as possible. It’s also ideal for legal professionals who are newer to their area of law, and need a subtle guiding hand. For all practices, ticklers essentially reduce the possible margin of error.
We spoke to Tad Thomas, a personal injury attorney practicing at Thomas Law Offices in Louisville, KY, about the benefits of using a case management program that embraces A.I. like Case Pacer on a daily basis. Tad Thomas speaks highly of the fact that Case Pacer, in particular, was created with plaintiff practices in mind.
Case Pacer gives my attorneys and I the tools we need to move plaintiff cases through the settlement process. The software’s computer learning elements give us the freedom we need to complete cases as quickly as possible without spending too much time on every step. This lets us spend more time with our clients, giving them the attention they deserve.
Search Engine Analytics
Every business that has an online presence utilizes the potential of search engine analytics, but law firms in particular benefit more than most other types of businesses. When potential clients need legal aid, they look for a local law firm that’s familiar with the specifics of their case. As soon as they find a firm willing to accept their case, they generally stop looking. This makes it integral for a firm to reach a fairly high position in their area’s local rankings. Ranking well in local searches can mean the difference between find new clientele and not.
Google’s search engines use hundreds of complex algorithms, but they use one in particular that utilizes computer learning in order to predict how users will search, why users will search, and what sites those users really want to view. This algorithm is called RankBrain. RankBrain knows what legal terminology frequently gets used and will predict what potential clients really mean when they type in terms they may not be familiar with.
It knows that a user, for example, who types in “lawyer accident injury in Chicago” most likely would like to get in touch with a personal injury lawyer in that city. RankBrain makes it easier for legal clients to find the lawyers they need. To accomplish this, good ol’ artificial intelligence lends a helping hand. To find out more, check out this article we wrote about RankBrain on Legal InSites.
To take advantage of complex search engine algorithms like RankBrain, lawyers who know how to properly use search engine optimization (SEO) techniques tend to rank better and receive more new clients. SEO professionals who are familiar with RankBrain, law firms, the legal industry as a whole, and the importance of local rankings are better prepared to help a law firm achieve their goals—both long and short term.
Wrapping it All Together—For the Future
Goofy headgear, bucket heads, and robotic voices are a far cry from the A.I. we’ve come to know, love, and use on a daily basis, but thanks to the A.I. we do have, we can all do our jobs easier and more efficient than ever before—letting us focus more on our families and on our clients.
During our interviews with each of the technology vendors and professionals above, we asked one question we thought was especially vital: How do you feel the legal industry will change with the assumed continuation of A.I. development? All of the answers we received to this question were profoundly positive and full of excitement. We didn’t have the space to publish all of those answers, unfortunately, but in short, they were inspiring.
The technology that surrounds us today is possible of tasks simply unimaginable. In ten—even five—brief years, it will be amazing to see where A.I. and computer learning take us. It’s an exciting time to be alive, and it’s an exciting time to be practicing law. Need evidence? Just take a look at this artificially intelligent lawyer. Pretty cool, right?
While the above list is a great starting point for legal professionals who are interested in A.I.-based platforms and products, it’s important to note that computer learning technology, in general, is still quite new. New software and new creations are being developed every day, all with one goal in mind—to pave a smarter, better path towards the future. A.I. technology lets us be more efficient than ever and help more clients than we thought possible. By embracing the tools and software of today, we guarantee a bright tomorrow.
Special thanks to Peter Wallqvist of RAVN Systems, Ned Gannon of eBrevia, Lori Brown of ComplianceHR, Keith Whitaker of Percipient, and Tad Thomas of Thomas Law Offices for speaking with us about their product offerings and experiences with A.I.-based programs and applications for law firms.
Follow Ryan Raplee on Twitter: https://twitter.com/ryanraplee